You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tapestry.apache.org by "Blower, Andy" <An...@proquest.co.uk> on 2010/04/20 16:16:11 UTC

Tapestry using 1.3Gb of heap space after capacity testing

We've been doing some capacity testing recently, and I'm concerned about the amount of heap memory that doesn't get freed when all the sessions expire. I'm seeing 1.3Gb after being left with no use overnight. I dumped the heap and took a look at what was using this much memory with the YourKit profiler. It looks to me like it's the Tapestry Page cache and javaassist stuff.

Here's a screenshot of the largest classes:

http://lh3.ggpht.com/_YwJn8TJTqJU/S82xSv1DPfI/AAAAAAAAABk/exlCHmPfy5E/T5.1-1300Mb-Heap.jpg


Has anyone else seen this and is this what should be expected with Tapestry?  It seems excessive to me, and leaves very little heap for our application to use, we currently have a 2Gb heap limit, and although I know this can be raised I'm still concerned.

Thanks,

Andy.


P.S. If there's anything else I should look at in the heap dump let me know, or if anyone wants the dump I can make it available. (300Mb zipped)


Re: Tapestry using 1.3Gb of heap space after capacity testing

Posted by Peter Stavrinides <P....@albourne.com>.
Hi Andy,

If you are running the defaults, then you might benefit from reducing the soft-limit of the pool, and increasing the hard limit, see this old thread: 
http://old.nabble.com/T5:-tapestry.page-pool.hard-limit-for-high-traffic-websites-td22971831.html

Cheers,
Peter


----- Original Message -----
From: "Thiago H. de Paula Figueiredo" <th...@gmail.com>
To: "Tapestry users" <us...@tapestry.apache.org>
Sent: Tuesday, 20 April, 2010 19:47:08 GMT +02:00 Athens, Beirut, Bucharest, Istanbul
Subject: Re: Tapestry using 1.3Gb of heap space after capacity testing

On Tue, 20 Apr 2010 13:41:37 -0300, Blower, Andy  
<An...@proquest.co.uk> wrote:

> It looks like explicit repeated use of a component, rather than using a  
> single component in a loop, creates a lot more component objects in the  
> heap.

Every <t:xxx> or <xxx t:type=""> is exactly one component, regardless of  
being inside a loop or not. When inside a loop, it's rendered many times,  
but instantiated just once, when the page instance is created. Static  
structure, dynamic behaviour. ;)

-- 
Thiago H. de Paula Figueiredo
Independent Java, Apache Tapestry 5 and Hibernate consultant, developer,  
and instructor
Owner, software architect and developer, Ars Machina Tecnologia da  
Informação Ltda.
http://www.arsmachina.com.br

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
For additional commands, e-mail: users-help@tapestry.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
For additional commands, e-mail: users-help@tapestry.apache.org


Re: Tapestry using 1.3Gb of heap space after capacity testing

Posted by "Thiago H. de Paula Figueiredo" <th...@gmail.com>.
On Tue, 20 Apr 2010 13:41:37 -0300, Blower, Andy  
<An...@proquest.co.uk> wrote:

> It looks like explicit repeated use of a component, rather than using a  
> single component in a loop, creates a lot more component objects in the  
> heap.

Every <t:xxx> or <xxx t:type=""> is exactly one component, regardless of  
being inside a loop or not. When inside a loop, it's rendered many times,  
but instantiated just once, when the page instance is created. Static  
structure, dynamic behaviour. ;)

-- 
Thiago H. de Paula Figueiredo
Independent Java, Apache Tapestry 5 and Hibernate consultant, developer,  
and instructor
Owner, software architect and developer, Ars Machina Tecnologia da  
Informação Ltda.
http://www.arsmachina.com.br

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
For additional commands, e-mail: users-help@tapestry.apache.org


Re: Tapestry using 1.3Gb of heap space after capacity testing

Posted by Pierce Wetter <pi...@paceap.com>.
> 
> 
> Anyone think I'm on the right track, or barking up the wrong tree completely?


  Here's a completely obscure question out of left field. Are you launching the JVM with -server?

 Here's a pretty useful blog post:

http://jfarcand.wordpress.com/2009/11/27/putting-glassfish-v3-in-production-essential-surviving-guide/

 There's lots of interesting meaty-but-obscure command line options to experiment with. 

 pierce
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
For additional commands, e-mail: users-help@tapestry.apache.org


Re: Tapestry using 1.3Gb of heap space after capacity testing

Posted by Howard Lewis Ship <hl...@gmail.com>.
On Wed, Jun 30, 2010 at 9:37 AM, Blower, Andy
<An...@proquest.co.uk> wrote:
>> -----Original Message-----
>> From: Howard Lewis Ship [mailto:hlship@gmail.com]
>> Sent: 29 June 2010 18:20
>> To: Tapestry users
>> Subject: Re: Tapestry using 1.3Gb of heap space after capacity testing
>>
>> It may be time to return to a more radical idea, one that is more
>> technically feasible now (in release 5.2) than it was in the past.
>
> Why more feasible in 5.2? We're using 5.1.0.5, and are now too close to release to upgrade. I was hoping to move to 5.2 when I thought it would be released in H1 2010, but we'll need to stick with 5.1 now I think. We launch in August.
>
>> Get rid of page pooling.
>>
>> I'm not saying to re-create each page for each request; I don't think
>> that would scale.
>>
>> However, it may be possible to change Tapestry so that you need only
>> ONE instance of a given page (per locale). All transient (per-request)
>> state for the page could be accessed indirectly, stored in a
>> per-thread object (or inside the Request, as attributes).
>>
>> I can envision some small amount of extra overhead per request (due to
>> extra levels of indirection).
>>
>> However, you could then process any number of threads against the
>> same, single page object with no extra memory consumption: for
>> example, no more duplicated Binding objects, and no more extra Maps to
>> hold them all.
>>
>> Some parts of the Tapestry's internal implementation
>> (InternalComponentResourcesImpl and ComponentPageElementImpl) would
>> need some changing, i.e., a bit more synchronization around some
>> critical sections.
>>
>> It's an exciting idea ... when will I have time to investigate it?
>
> It certainly is an exciting idea, out of curiosity How long do you think it would take you to implement this - it sounds like a pretty big overhaul to me.


It's not so large an overhaul as you might think. My big concern is
that the change in semantics (not the change in API) might cause
existing ComponentClassTransformationWorkers (contributed by 3rd party
libraries) to no longer operate correctly.

>
> Anything else that I could look into? Am I on the right track?
>
>
>> -----Original Message-----
>> From: Christophe Cordenier [mailto:christophe.cordenier@gmail.com]
>> Sent: 29 June 2010 19:35
>> To: Tapestry users
>> Subject: Re: Tapestry using 1.3Gb of heap space after capacity testing
>>
>> Holy grail ! but is it really feasible...
>>
>> @Andy Sorry i didn't follow all the thread but if i understand well, you
>> have page instance of 7.5Mb for one single instance ? This is only due to
>> page and component structure ?
>
> That's correct I believe - it used to be 14Mb. Our largest page is 10Mb and we have 13 pages with over 4Mb retained size from their PageImpl's.
>
> I'm not sure how this compares with the average, but I suspect they are a bit on the large size.
>
> Andy
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
> For additional commands, e-mail: users-help@tapestry.apache.org
>
>



-- 
Howard M. Lewis Ship

Creator of Apache Tapestry

The source for Tapestry training, mentoring and support. Contact me to
learn how I can get you up and productive in Tapestry fast!

(971) 678-5210
http://howardlewisship.com

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
For additional commands, e-mail: users-help@tapestry.apache.org


Re: Tapestry using 1.3Gb of heap space after capacity testing

Posted by "Thiago H. de Paula Figueiredo" <th...@gmail.com>.
On Wed, 30 Jun 2010 13:37:54 -0300, Blower, Andy  
<An...@proquest.co.uk> wrote:

> Why more feasible in 5.2? We're using 5.1.0.5, and are now too close to  
> release to upgrade. I was hoping to move to 5.2 when I thought it would  
> be released in H1 2010, but we'll need to stick with 5.1 now I think. We  
> launch in August.

I guess because of the changes on how class transformations are done.

-- 
Thiago H. de Paula Figueiredo
Independent Java, Apache Tapestry 5 and Hibernate consultant, developer,  
and instructor
Owner, Ars Machina Tecnologia da Informação Ltda.
http://www.arsmachina.com.br

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
For additional commands, e-mail: users-help@tapestry.apache.org


RE: Tapestry using 1.3Gb of heap space after capacity testing

Posted by "Blower, Andy" <An...@proquest.co.uk>.
> -----Original Message-----
> From: Howard Lewis Ship [mailto:hlship@gmail.com]
> Sent: 29 June 2010 18:20
> To: Tapestry users
> Subject: Re: Tapestry using 1.3Gb of heap space after capacity testing
> 
> It may be time to return to a more radical idea, one that is more
> technically feasible now (in release 5.2) than it was in the past.

Why more feasible in 5.2? We're using 5.1.0.5, and are now too close to release to upgrade. I was hoping to move to 5.2 when I thought it would be released in H1 2010, but we'll need to stick with 5.1 now I think. We launch in August.
 
> Get rid of page pooling.
> 
> I'm not saying to re-create each page for each request; I don't think
> that would scale.
> 
> However, it may be possible to change Tapestry so that you need only
> ONE instance of a given page (per locale). All transient (per-request)
> state for the page could be accessed indirectly, stored in a
> per-thread object (or inside the Request, as attributes).
> 
> I can envision some small amount of extra overhead per request (due to
> extra levels of indirection).
> 
> However, you could then process any number of threads against the
> same, single page object with no extra memory consumption: for
> example, no more duplicated Binding objects, and no more extra Maps to
> hold them all.
> 
> Some parts of the Tapestry's internal implementation
> (InternalComponentResourcesImpl and ComponentPageElementImpl) would
> need some changing, i.e., a bit more synchronization around some
> critical sections.
> 
> It's an exciting idea ... when will I have time to investigate it?

It certainly is an exciting idea, out of curiosity How long do you think it would take you to implement this - it sounds like a pretty big overhaul to me.

Anything else that I could look into? Am I on the right track?


> -----Original Message-----
> From: Christophe Cordenier [mailto:christophe.cordenier@gmail.com]
> Sent: 29 June 2010 19:35
> To: Tapestry users
> Subject: Re: Tapestry using 1.3Gb of heap space after capacity testing
> 
> Holy grail ! but is it really feasible...
> 
> @Andy Sorry i didn't follow all the thread but if i understand well, you
> have page instance of 7.5Mb for one single instance ? This is only due to
> page and component structure ?

That's correct I believe - it used to be 14Mb. Our largest page is 10Mb and we have 13 pages with over 4Mb retained size from their PageImpl's.

I'm not sure how this compares with the average, but I suspect they are a bit on the large size.

Andy


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
For additional commands, e-mail: users-help@tapestry.apache.org


Re: Tapestry using 1.3Gb of heap space after capacity testing

Posted by Christophe Cordenier <ch...@gmail.com>.
Holy grail ! but is it really feasible...

@Andy Sorry i didn't follow all the thread but if i understand well, you
have page instance of 7.5Mb for one single instance ? This is only due to
page and component structure ?

2010/6/29 Howard Lewis Ship <hl...@gmail.com>

> It may be time to return to a more radical idea, one that is more
> technically feasible now (in release 5.2) than it was in the past.
>
> Get rid of page pooling.
>
> I'm not saying to re-create each page for each request; I don't think
> that would scale.
>
> However, it may be possible to change Tapestry so that you need only
> ONE instance of a given page (per locale). All transient (per-request)
> state for the page could be accessed indirectly, stored in a
> per-thread object (or inside the Request, as attributes).
>
> I can envision some small amount of extra overhead per request (due to
> extra levels of indirection).
>
> However, you could then process any number of threads against the
> same, single page object with no extra memory consumption: for
> example, no more duplicated Binding objects, and no more extra Maps to
> hold them all.
>
> Some parts of the Tapestry's internal implementation
> (InternalComponentResourcesImpl and ComponentPageElementImpl) would
> need some changing, i.e., a bit more synchronization around some
> critical sections.
>
> It's an exciting idea ... when will I have time to investigate it?
>
> On Tue, Jun 29, 2010 at 10:02 AM, Howard Lewis Ship <hl...@gmail.com>
> wrote:
> > Just to check ... which version of Tapestry?
> >
> > On Tue, Jun 29, 2010 at 8:48 AM, Blower, Andy
> > <An...@proquest.co.uk> wrote:
> >> I doubt anyone remembers this thread except me, but we're still having
> problems that I could do with help on.
> >>
> >> So, two months later, we have managed to improve things significantly,
> but still have very large pages and we're having a lot of trouble with the
> page pool size. What we see when we push the app too hard is that the heap
> fills until the gc thread is permanently running, and the app is essentially
> dead and unresponsive. I've been testing with 2g & 4g heaps up to now, with
> 100ms soft-wait & 200 hard limit because we need a lot of instances of some
> pages, and were running out with a 100 limit. I know I need to understand
> and do more with these values, but it's worrying because our tests are only
> requesting English pages at present and we'll have 17 languages for launch!
> So I'm rather concerned to say the least...
> >>
> >> I have a theory about what is happening that I'd like to run by the list
> and see what others think. I'm having some trouble figuring out what's
> happening at load.
> >>
> >> What I think is that when the heap gets full-ish at high load, the gc
> kicks in more and more often, slowing request processing down a little.
> (there's a fair amount of soft referenced stuff it can boot) This means that
> pages are returned to the page pool a bit later causing more page instances
> to be created which fills the heap and enters an vicious cycle with the gc
> trying to free memory and the T5 page pool trying to create more page
> instances. The reason I suspect this is because heap dumps show about 40
> PageImpl's that have not completed loading, and many more that have but are
> much smaller in size than others for the same page. (e.g. a Results PageImpl
> retaining 28k when a normal one retains 7.5Mb - very big I know, but I've
> cut it down from 14Mb)
> >>
> >> Our site has 223 pages which consume 213Mb if 1 instance of each is
> instantiated. Multiply this by 200 (hard limit) and then by 17 languages and
> worst case we'll need a heap size of 725Gb which is a little ridiculous! I
> will do more work on reducing the size of the pages but I've already done
> the easy stuff.
> >>
> >> Here's a list of the heap dominators using Eclipse MAT,
> http://lh6.ggpht.com/_YwJn8TJTqJU/TCoNT11ouZI/AAAAAAAAACI/G8eTGsth4zM/HeapDominators.png
> >>
> >> Anyone think I'm on the right track, or barking up the wrong tree
> completely?
> >>
> >> Thanks,
> >>
> >> Andy
> >>
> >>> -----Original Message-----
> >>> From: Blower, Andy [mailto:Andy.Blower@proquest.co.uk]
> >>> Sent: 20 April 2010 17:42
> >>> To: 'Tapestry users'
> >>> Subject: RE: Tapestry using 1.3Gb of heap space after capacity testing
> >>>
> >>> Thanks for the link. As I said before, I don't think that we've fallen
> >>> prey to that exactly. Doing some local testing and looking at one of
> >>> our largest pages which displays search results, it has 76 component
> >>> objects, each of those will have components nested within (pretty
> >>> heavily in some cases) but nothing which has a bunch of components
> >>> where only one is used and the others are redundant. (what I think is
> >>> this uber-component thing)
> >>>
> >>> It looks like explicit repeated use of a component, rather than using a
> >>> single component in a loop, creates a lot more component objects in the
> >>> heap. I've never seen anything warning against this, and we do usually
> >>> use loops but it's not always the best solution.
> >>>
> >>> Again, looking at our results page, a single instance seems to get
> >>> 11,283 ComponentPageElementImpl instances created for it. I'm finding
> >>> it hard to find a good view of the tree to figure out what component(s)
> >>> are the main culprits for this huge number of instances.
> >>>
> >>> Thanks for any help or guidance you can give me.
> >>>
> >>> Andy
> >>>
> >>> > -----Original Message-----
> >>> > From: Thiago H. de Paula Figueiredo [mailto:thiagohp@gmail.com]
> >>> > Sent: 20 April 2010 15:56
> >>> > To: Tapestry users
> >>> > Subject: Re: Tapestry using 1.3Gb of heap space after capacity
> >>> testing
> >>> >
> >>> > On Tue, 20 Apr 2010 11:41:25 -0300, Blower, Andy
> >>> > <An...@proquest.co.uk> wrote:
> >>> >
> >>> > > Uber-component anti-pattern, not to my knowledge, but we have a lot
> >>> > of
> >>> > > pages and components now. And a lot of logic in them. The project's
> >>> > > large enough I can't be sure. Do you have a handy link to Howard's
> >>> > > description so I can get my team to check?
> >>> >
> >>> > It's here:
> >>> >
> http://old.nabble.com/-T5.0.18--Out-of-Memory-Error---Potential-Leak-
> >>> > (doesn't-reduce-after-forced-GC)-to25403474s302.html#a25497441
> >>> > That thread is interesting, by the way. The problem is more about how
> >>> > components and pages are written than the number of them.
> >>> > What Tapestry version are you using?
> >>> >
> >>> > > Page pool config is default, I don't know enough to fiddle yet.
> >>> Also,
> >>> > > the test is on a single stand alone server with no clustering, we
> >>> are
> >>> > > using Tomcat clustering for production.
> >>> >
> >>> > You should take a look a them, as they can affect the memory
> >>> > consumption
> >>> > directly.
> >>> >
> >>> > > I did fire the GC manually before getting concerned BTW. A lot of
> >>> > heap
> >>> > > seems to be taken up with 100-300k descriptions on the
> >>> > > InternalClassTransformationImpl class. An example of one is
> >>> attached
> >>> > > below - seems wrong to me and more like the debug that can be
> >>> > requested
> >>> > > through log4j rather than a description.
> >>> >
> >>> > They look really bigger than they should. 300k is a lot. I have one
> >>> > application with a reasonable number of pages and components and it
> >>> > runs
> >>> > inside a 512 MB slide at SliceHost without using the whole memory.
> >>> >
> >>> > --
> >>> > Thiago H. de Paula Figueiredo
> >>> > Independent Java, Apache Tapestry 5 and Hibernate consultant,
> >>> > developer,
> >>> > and instructor
> >>> > Owner, software architect and developer, Ars Machina Tecnologia da
> >>> > Informação Ltda.
> >>> > http://www.arsmachina.com.br
> >>> >
> >>> > ---------------------------------------------------------------------
> >>> > To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
> >>> > For additional commands, e-mail: users-help@tapestry.apache.org
> >>> >
> >>
> >>
> >
> >
> >
> > --
> > Howard M. Lewis Ship
> >
> > Creator of Apache Tapestry
> >
> > The source for Tapestry training, mentoring and support. Contact me to
> > learn how I can get you up and productive in Tapestry fast!
> >
> > (971) 678-5210
> > http://howardlewisship.com
> >
>
>
>
> --
> Howard M. Lewis Ship
>
> Creator of Apache Tapestry
>
> The source for Tapestry training, mentoring and support. Contact me to
> learn how I can get you up and productive in Tapestry fast!
>
> (971) 678-5210
> http://howardlewisship.com
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
> For additional commands, e-mail: users-help@tapestry.apache.org
>
>


-- 
Regards,
Christophe Cordenier.

Committer on Apache Tapestry 5
Co-creator of wooki @wookicentral.com

Re: Tapestry using 1.3Gb of heap space after capacity testing

Posted by Howard Lewis Ship <hl...@gmail.com>.
It may be time to return to a more radical idea, one that is more
technically feasible now (in release 5.2) than it was in the past.

Get rid of page pooling.

I'm not saying to re-create each page for each request; I don't think
that would scale.

However, it may be possible to change Tapestry so that you need only
ONE instance of a given page (per locale). All transient (per-request)
state for the page could be accessed indirectly, stored in a
per-thread object (or inside the Request, as attributes).

I can envision some small amount of extra overhead per request (due to
extra levels of indirection).

However, you could then process any number of threads against the
same, single page object with no extra memory consumption: for
example, no more duplicated Binding objects, and no more extra Maps to
hold them all.

Some parts of the Tapestry's internal implementation
(InternalComponentResourcesImpl and ComponentPageElementImpl) would
need some changing, i.e., a bit more synchronization around some
critical sections.

It's an exciting idea ... when will I have time to investigate it?

On Tue, Jun 29, 2010 at 10:02 AM, Howard Lewis Ship <hl...@gmail.com> wrote:
> Just to check ... which version of Tapestry?
>
> On Tue, Jun 29, 2010 at 8:48 AM, Blower, Andy
> <An...@proquest.co.uk> wrote:
>> I doubt anyone remembers this thread except me, but we're still having problems that I could do with help on.
>>
>> So, two months later, we have managed to improve things significantly, but still have very large pages and we're having a lot of trouble with the page pool size. What we see when we push the app too hard is that the heap fills until the gc thread is permanently running, and the app is essentially dead and unresponsive. I've been testing with 2g & 4g heaps up to now, with 100ms soft-wait & 200 hard limit because we need a lot of instances of some pages, and were running out with a 100 limit. I know I need to understand and do more with these values, but it's worrying because our tests are only requesting English pages at present and we'll have 17 languages for launch! So I'm rather concerned to say the least...
>>
>> I have a theory about what is happening that I'd like to run by the list and see what others think. I'm having some trouble figuring out what's happening at load.
>>
>> What I think is that when the heap gets full-ish at high load, the gc kicks in more and more often, slowing request processing down a little. (there's a fair amount of soft referenced stuff it can boot) This means that pages are returned to the page pool a bit later causing more page instances to be created which fills the heap and enters an vicious cycle with the gc trying to free memory and the T5 page pool trying to create more page instances. The reason I suspect this is because heap dumps show about 40 PageImpl's that have not completed loading, and many more that have but are much smaller in size than others for the same page. (e.g. a Results PageImpl retaining 28k when a normal one retains 7.5Mb - very big I know, but I've cut it down from 14Mb)
>>
>> Our site has 223 pages which consume 213Mb if 1 instance of each is instantiated. Multiply this by 200 (hard limit) and then by 17 languages and worst case we'll need a heap size of 725Gb which is a little ridiculous! I will do more work on reducing the size of the pages but I've already done the easy stuff.
>>
>> Here's a list of the heap dominators using Eclipse MAT, http://lh6.ggpht.com/_YwJn8TJTqJU/TCoNT11ouZI/AAAAAAAAACI/G8eTGsth4zM/HeapDominators.png
>>
>> Anyone think I'm on the right track, or barking up the wrong tree completely?
>>
>> Thanks,
>>
>> Andy
>>
>>> -----Original Message-----
>>> From: Blower, Andy [mailto:Andy.Blower@proquest.co.uk]
>>> Sent: 20 April 2010 17:42
>>> To: 'Tapestry users'
>>> Subject: RE: Tapestry using 1.3Gb of heap space after capacity testing
>>>
>>> Thanks for the link. As I said before, I don't think that we've fallen
>>> prey to that exactly. Doing some local testing and looking at one of
>>> our largest pages which displays search results, it has 76 component
>>> objects, each of those will have components nested within (pretty
>>> heavily in some cases) but nothing which has a bunch of components
>>> where only one is used and the others are redundant. (what I think is
>>> this uber-component thing)
>>>
>>> It looks like explicit repeated use of a component, rather than using a
>>> single component in a loop, creates a lot more component objects in the
>>> heap. I've never seen anything warning against this, and we do usually
>>> use loops but it's not always the best solution.
>>>
>>> Again, looking at our results page, a single instance seems to get
>>> 11,283 ComponentPageElementImpl instances created for it. I'm finding
>>> it hard to find a good view of the tree to figure out what component(s)
>>> are the main culprits for this huge number of instances.
>>>
>>> Thanks for any help or guidance you can give me.
>>>
>>> Andy
>>>
>>> > -----Original Message-----
>>> > From: Thiago H. de Paula Figueiredo [mailto:thiagohp@gmail.com]
>>> > Sent: 20 April 2010 15:56
>>> > To: Tapestry users
>>> > Subject: Re: Tapestry using 1.3Gb of heap space after capacity
>>> testing
>>> >
>>> > On Tue, 20 Apr 2010 11:41:25 -0300, Blower, Andy
>>> > <An...@proquest.co.uk> wrote:
>>> >
>>> > > Uber-component anti-pattern, not to my knowledge, but we have a lot
>>> > of
>>> > > pages and components now. And a lot of logic in them. The project's
>>> > > large enough I can't be sure. Do you have a handy link to Howard's
>>> > > description so I can get my team to check?
>>> >
>>> > It's here:
>>> > http://old.nabble.com/-T5.0.18--Out-of-Memory-Error---Potential-Leak-
>>> > (doesn't-reduce-after-forced-GC)-to25403474s302.html#a25497441
>>> > That thread is interesting, by the way. The problem is more about how
>>> > components and pages are written than the number of them.
>>> > What Tapestry version are you using?
>>> >
>>> > > Page pool config is default, I don't know enough to fiddle yet.
>>> Also,
>>> > > the test is on a single stand alone server with no clustering, we
>>> are
>>> > > using Tomcat clustering for production.
>>> >
>>> > You should take a look a them, as they can affect the memory
>>> > consumption
>>> > directly.
>>> >
>>> > > I did fire the GC manually before getting concerned BTW. A lot of
>>> > heap
>>> > > seems to be taken up with 100-300k descriptions on the
>>> > > InternalClassTransformationImpl class. An example of one is
>>> attached
>>> > > below - seems wrong to me and more like the debug that can be
>>> > requested
>>> > > through log4j rather than a description.
>>> >
>>> > They look really bigger than they should. 300k is a lot. I have one
>>> > application with a reasonable number of pages and components and it
>>> > runs
>>> > inside a 512 MB slide at SliceHost without using the whole memory.
>>> >
>>> > --
>>> > Thiago H. de Paula Figueiredo
>>> > Independent Java, Apache Tapestry 5 and Hibernate consultant,
>>> > developer,
>>> > and instructor
>>> > Owner, software architect and developer, Ars Machina Tecnologia da
>>> > Informação Ltda.
>>> > http://www.arsmachina.com.br
>>> >
>>> > ---------------------------------------------------------------------
>>> > To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
>>> > For additional commands, e-mail: users-help@tapestry.apache.org
>>> >
>>
>>
>
>
>
> --
> Howard M. Lewis Ship
>
> Creator of Apache Tapestry
>
> The source for Tapestry training, mentoring and support. Contact me to
> learn how I can get you up and productive in Tapestry fast!
>
> (971) 678-5210
> http://howardlewisship.com
>



-- 
Howard M. Lewis Ship

Creator of Apache Tapestry

The source for Tapestry training, mentoring and support. Contact me to
learn how I can get you up and productive in Tapestry fast!

(971) 678-5210
http://howardlewisship.com

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
For additional commands, e-mail: users-help@tapestry.apache.org


Re: Tapestry using 1.3Gb of heap space after capacity testing

Posted by Howard Lewis Ship <hl...@gmail.com>.
Just to check ... which version of Tapestry?

On Tue, Jun 29, 2010 at 8:48 AM, Blower, Andy
<An...@proquest.co.uk> wrote:
> I doubt anyone remembers this thread except me, but we're still having problems that I could do with help on.
>
> So, two months later, we have managed to improve things significantly, but still have very large pages and we're having a lot of trouble with the page pool size. What we see when we push the app too hard is that the heap fills until the gc thread is permanently running, and the app is essentially dead and unresponsive. I've been testing with 2g & 4g heaps up to now, with 100ms soft-wait & 200 hard limit because we need a lot of instances of some pages, and were running out with a 100 limit. I know I need to understand and do more with these values, but it's worrying because our tests are only requesting English pages at present and we'll have 17 languages for launch! So I'm rather concerned to say the least...
>
> I have a theory about what is happening that I'd like to run by the list and see what others think. I'm having some trouble figuring out what's happening at load.
>
> What I think is that when the heap gets full-ish at high load, the gc kicks in more and more often, slowing request processing down a little. (there's a fair amount of soft referenced stuff it can boot) This means that pages are returned to the page pool a bit later causing more page instances to be created which fills the heap and enters an vicious cycle with the gc trying to free memory and the T5 page pool trying to create more page instances. The reason I suspect this is because heap dumps show about 40 PageImpl's that have not completed loading, and many more that have but are much smaller in size than others for the same page. (e.g. a Results PageImpl retaining 28k when a normal one retains 7.5Mb - very big I know, but I've cut it down from 14Mb)
>
> Our site has 223 pages which consume 213Mb if 1 instance of each is instantiated. Multiply this by 200 (hard limit) and then by 17 languages and worst case we'll need a heap size of 725Gb which is a little ridiculous! I will do more work on reducing the size of the pages but I've already done the easy stuff.
>
> Here's a list of the heap dominators using Eclipse MAT, http://lh6.ggpht.com/_YwJn8TJTqJU/TCoNT11ouZI/AAAAAAAAACI/G8eTGsth4zM/HeapDominators.png
>
> Anyone think I'm on the right track, or barking up the wrong tree completely?
>
> Thanks,
>
> Andy
>
>> -----Original Message-----
>> From: Blower, Andy [mailto:Andy.Blower@proquest.co.uk]
>> Sent: 20 April 2010 17:42
>> To: 'Tapestry users'
>> Subject: RE: Tapestry using 1.3Gb of heap space after capacity testing
>>
>> Thanks for the link. As I said before, I don't think that we've fallen
>> prey to that exactly. Doing some local testing and looking at one of
>> our largest pages which displays search results, it has 76 component
>> objects, each of those will have components nested within (pretty
>> heavily in some cases) but nothing which has a bunch of components
>> where only one is used and the others are redundant. (what I think is
>> this uber-component thing)
>>
>> It looks like explicit repeated use of a component, rather than using a
>> single component in a loop, creates a lot more component objects in the
>> heap. I've never seen anything warning against this, and we do usually
>> use loops but it's not always the best solution.
>>
>> Again, looking at our results page, a single instance seems to get
>> 11,283 ComponentPageElementImpl instances created for it. I'm finding
>> it hard to find a good view of the tree to figure out what component(s)
>> are the main culprits for this huge number of instances.
>>
>> Thanks for any help or guidance you can give me.
>>
>> Andy
>>
>> > -----Original Message-----
>> > From: Thiago H. de Paula Figueiredo [mailto:thiagohp@gmail.com]
>> > Sent: 20 April 2010 15:56
>> > To: Tapestry users
>> > Subject: Re: Tapestry using 1.3Gb of heap space after capacity
>> testing
>> >
>> > On Tue, 20 Apr 2010 11:41:25 -0300, Blower, Andy
>> > <An...@proquest.co.uk> wrote:
>> >
>> > > Uber-component anti-pattern, not to my knowledge, but we have a lot
>> > of
>> > > pages and components now. And a lot of logic in them. The project's
>> > > large enough I can't be sure. Do you have a handy link to Howard's
>> > > description so I can get my team to check?
>> >
>> > It's here:
>> > http://old.nabble.com/-T5.0.18--Out-of-Memory-Error---Potential-Leak-
>> > (doesn't-reduce-after-forced-GC)-to25403474s302.html#a25497441
>> > That thread is interesting, by the way. The problem is more about how
>> > components and pages are written than the number of them.
>> > What Tapestry version are you using?
>> >
>> > > Page pool config is default, I don't know enough to fiddle yet.
>> Also,
>> > > the test is on a single stand alone server with no clustering, we
>> are
>> > > using Tomcat clustering for production.
>> >
>> > You should take a look a them, as they can affect the memory
>> > consumption
>> > directly.
>> >
>> > > I did fire the GC manually before getting concerned BTW. A lot of
>> > heap
>> > > seems to be taken up with 100-300k descriptions on the
>> > > InternalClassTransformationImpl class. An example of one is
>> attached
>> > > below - seems wrong to me and more like the debug that can be
>> > requested
>> > > through log4j rather than a description.
>> >
>> > They look really bigger than they should. 300k is a lot. I have one
>> > application with a reasonable number of pages and components and it
>> > runs
>> > inside a 512 MB slide at SliceHost without using the whole memory.
>> >
>> > --
>> > Thiago H. de Paula Figueiredo
>> > Independent Java, Apache Tapestry 5 and Hibernate consultant,
>> > developer,
>> > and instructor
>> > Owner, software architect and developer, Ars Machina Tecnologia da
>> > Informação Ltda.
>> > http://www.arsmachina.com.br
>> >
>> > ---------------------------------------------------------------------
>> > To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
>> > For additional commands, e-mail: users-help@tapestry.apache.org
>> >
>
>



-- 
Howard M. Lewis Ship

Creator of Apache Tapestry

The source for Tapestry training, mentoring and support. Contact me to
learn how I can get you up and productive in Tapestry fast!

(971) 678-5210
http://howardlewisship.com

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
For additional commands, e-mail: users-help@tapestry.apache.org


Re: Tapestry using 1.3Gb of heap space after capacity testing

Posted by Christophe Cordenier <ch...@gmail.com>.
Hi

This make me think about something i have found when i started to go deep in
the Tapestry page pool code. It's about the PagePoolImpl service. If we look
at the checkout method, we can see that the test made to check the pool's
limits is made on a inUse list. I have made the experience that in heavy
load when the pool is empty this test becomes useless because the inUse list
is updated after the page loading (that is obviously not synchronized).
Theory is that if you have huge page, you may come into this case.

I don't know if it's related to your problem, but I guess, this is why
"tapestry.page-pool.hard-limit" will be deprecated, as this is to the
container to handle the limits of request handling threads.

Anyway, in the meantime shouldn't we handle the limits check using a
'inConstruction' list ? But it sounds to simple to be true :)


2010/6/29 Blower, Andy <An...@proquest.co.uk>

> I doubt anyone remembers this thread except me, but we're still having
> problems that I could do with help on.
>
> So, two months later, we have managed to improve things significantly, but
> still have very large pages and we're having a lot of trouble with the page
> pool size. What we see when we push the app too hard is that the heap fills
> until the gc thread is permanently running, and the app is essentially dead
> and unresponsive. I've been testing with 2g & 4g heaps up to now, with 100ms
> soft-wait & 200 hard limit because we need a lot of instances of some pages,
> and were running out with a 100 limit. I know I need to understand and do
> more with these values, but it's worrying because our tests are only
> requesting English pages at present and we'll have 17 languages for launch!
> So I'm rather concerned to say the least...
>
> I have a theory about what is happening that I'd like to run by the list
> and see what others think. I'm having some trouble figuring out what's
> happening at load.
>
> What I think is that when the heap gets full-ish at high load, the gc kicks
> in more and more often, slowing request processing down a little. (there's a
> fair amount of soft referenced stuff it can boot) This means that pages are
> returned to the page pool a bit later causing more page instances to be
> created which fills the heap and enters an vicious cycle with the gc trying
> to free memory and the T5 page pool trying to create more page instances.
> The reason I suspect this is because heap dumps show about 40 PageImpl's
> that have not completed loading, and many more that have but are much
> smaller in size than others for the same page. (e.g. a Results PageImpl
> retaining 28k when a normal one retains 7.5Mb - very big I know, but I've
> cut it down from 14Mb)
>
> Our site has 223 pages which consume 213Mb if 1 instance of each is
> instantiated. Multiply this by 200 (hard limit) and then by 17 languages and
> worst case we'll need a heap size of 725Gb which is a little ridiculous! I
> will do more work on reducing the size of the pages but I've already done
> the easy stuff.
>
> Here's a list of the heap dominators using Eclipse MAT,
> http://lh6.ggpht.com/_YwJn8TJTqJU/TCoNT11ouZI/AAAAAAAAACI/G8eTGsth4zM/HeapDominators.png
>
> Anyone think I'm on the right track, or barking up the wrong tree
> completely?
>
> Thanks,
>
> Andy
>
> > -----Original Message-----
> > From: Blower, Andy [mailto:Andy.Blower@proquest.co.uk]
> > Sent: 20 April 2010 17:42
> > To: 'Tapestry users'
> > Subject: RE: Tapestry using 1.3Gb of heap space after capacity testing
> >
> > Thanks for the link. As I said before, I don't think that we've fallen
> > prey to that exactly. Doing some local testing and looking at one of
> > our largest pages which displays search results, it has 76 component
> > objects, each of those will have components nested within (pretty
> > heavily in some cases) but nothing which has a bunch of components
> > where only one is used and the others are redundant. (what I think is
> > this uber-component thing)
> >
> > It looks like explicit repeated use of a component, rather than using a
> > single component in a loop, creates a lot more component objects in the
> > heap. I've never seen anything warning against this, and we do usually
> > use loops but it's not always the best solution.
> >
> > Again, looking at our results page, a single instance seems to get
> > 11,283 ComponentPageElementImpl instances created for it. I'm finding
> > it hard to find a good view of the tree to figure out what component(s)
> > are the main culprits for this huge number of instances.
> >
> > Thanks for any help or guidance you can give me.
> >
> > Andy
> >
> > > -----Original Message-----
> > > From: Thiago H. de Paula Figueiredo [mailto:thiagohp@gmail.com]
> > > Sent: 20 April 2010 15:56
> > > To: Tapestry users
> > > Subject: Re: Tapestry using 1.3Gb of heap space after capacity
> > testing
> > >
> > > On Tue, 20 Apr 2010 11:41:25 -0300, Blower, Andy
> > > <An...@proquest.co.uk> wrote:
> > >
> > > > Uber-component anti-pattern, not to my knowledge, but we have a lot
> > > of
> > > > pages and components now. And a lot of logic in them. The project's
> > > > large enough I can't be sure. Do you have a handy link to Howard's
> > > > description so I can get my team to check?
> > >
> > > It's here:
> > > http://old.nabble.com/-T5.0.18--Out-of-Memory-Error---Potential-Leak-
> > > (doesn't-reduce-after-forced-GC)-to25403474s302.html#a25497441
> > > That thread is interesting, by the way. The problem is more about how
> > > components and pages are written than the number of them.
> > > What Tapestry version are you using?
> > >
> > > > Page pool config is default, I don't know enough to fiddle yet.
> > Also,
> > > > the test is on a single stand alone server with no clustering, we
> > are
> > > > using Tomcat clustering for production.
> > >
> > > You should take a look a them, as they can affect the memory
> > > consumption
> > > directly.
> > >
> > > > I did fire the GC manually before getting concerned BTW. A lot of
> > > heap
> > > > seems to be taken up with 100-300k descriptions on the
> > > > InternalClassTransformationImpl class. An example of one is
> > attached
> > > > below - seems wrong to me and more like the debug that can be
> > > requested
> > > > through log4j rather than a description.
> > >
> > > They look really bigger than they should. 300k is a lot. I have one
> > > application with a reasonable number of pages and components and it
> > > runs
> > > inside a 512 MB slide at SliceHost without using the whole memory.
> > >
> > > --
> > > Thiago H. de Paula Figueiredo
> > > Independent Java, Apache Tapestry 5 and Hibernate consultant,
> > > developer,
> > > and instructor
> > > Owner, software architect and developer, Ars Machina Tecnologia da
> > > Informação Ltda.
> > > http://www.arsmachina.com.br
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
> > > For additional commands, e-mail: users-help@tapestry.apache.org
> > >
>
>


-- 
Regards,
Christophe Cordenier.

Committer on Apache Tapestry 5
Co-creator of wooki @wookicentral.com

RE: Tapestry using 1.3Gb of heap space after capacity testing

Posted by "Blower, Andy" <An...@proquest.co.uk>.
I doubt anyone remembers this thread except me, but we're still having problems that I could do with help on.

So, two months later, we have managed to improve things significantly, but still have very large pages and we're having a lot of trouble with the page pool size. What we see when we push the app too hard is that the heap fills until the gc thread is permanently running, and the app is essentially dead and unresponsive. I've been testing with 2g & 4g heaps up to now, with 100ms soft-wait & 200 hard limit because we need a lot of instances of some pages, and were running out with a 100 limit. I know I need to understand and do more with these values, but it's worrying because our tests are only requesting English pages at present and we'll have 17 languages for launch! So I'm rather concerned to say the least...

I have a theory about what is happening that I'd like to run by the list and see what others think. I'm having some trouble figuring out what's happening at load.

What I think is that when the heap gets full-ish at high load, the gc kicks in more and more often, slowing request processing down a little. (there's a fair amount of soft referenced stuff it can boot) This means that pages are returned to the page pool a bit later causing more page instances to be created which fills the heap and enters an vicious cycle with the gc trying to free memory and the T5 page pool trying to create more page instances. The reason I suspect this is because heap dumps show about 40 PageImpl's that have not completed loading, and many more that have but are much smaller in size than others for the same page. (e.g. a Results PageImpl retaining 28k when a normal one retains 7.5Mb - very big I know, but I've cut it down from 14Mb) 

Our site has 223 pages which consume 213Mb if 1 instance of each is instantiated. Multiply this by 200 (hard limit) and then by 17 languages and worst case we'll need a heap size of 725Gb which is a little ridiculous! I will do more work on reducing the size of the pages but I've already done the easy stuff.

Here's a list of the heap dominators using Eclipse MAT, http://lh6.ggpht.com/_YwJn8TJTqJU/TCoNT11ouZI/AAAAAAAAACI/G8eTGsth4zM/HeapDominators.png

Anyone think I'm on the right track, or barking up the wrong tree completely?

Thanks,

Andy

> -----Original Message-----
> From: Blower, Andy [mailto:Andy.Blower@proquest.co.uk]
> Sent: 20 April 2010 17:42
> To: 'Tapestry users'
> Subject: RE: Tapestry using 1.3Gb of heap space after capacity testing
> 
> Thanks for the link. As I said before, I don't think that we've fallen
> prey to that exactly. Doing some local testing and looking at one of
> our largest pages which displays search results, it has 76 component
> objects, each of those will have components nested within (pretty
> heavily in some cases) but nothing which has a bunch of components
> where only one is used and the others are redundant. (what I think is
> this uber-component thing)
> 
> It looks like explicit repeated use of a component, rather than using a
> single component in a loop, creates a lot more component objects in the
> heap. I've never seen anything warning against this, and we do usually
> use loops but it's not always the best solution.
> 
> Again, looking at our results page, a single instance seems to get
> 11,283 ComponentPageElementImpl instances created for it. I'm finding
> it hard to find a good view of the tree to figure out what component(s)
> are the main culprits for this huge number of instances.
> 
> Thanks for any help or guidance you can give me.
> 
> Andy
> 
> > -----Original Message-----
> > From: Thiago H. de Paula Figueiredo [mailto:thiagohp@gmail.com]
> > Sent: 20 April 2010 15:56
> > To: Tapestry users
> > Subject: Re: Tapestry using 1.3Gb of heap space after capacity
> testing
> >
> > On Tue, 20 Apr 2010 11:41:25 -0300, Blower, Andy
> > <An...@proquest.co.uk> wrote:
> >
> > > Uber-component anti-pattern, not to my knowledge, but we have a lot
> > of
> > > pages and components now. And a lot of logic in them. The project's
> > > large enough I can't be sure. Do you have a handy link to Howard's
> > > description so I can get my team to check?
> >
> > It's here:
> > http://old.nabble.com/-T5.0.18--Out-of-Memory-Error---Potential-Leak-
> > (doesn't-reduce-after-forced-GC)-to25403474s302.html#a25497441
> > That thread is interesting, by the way. The problem is more about how
> > components and pages are written than the number of them.
> > What Tapestry version are you using?
> >
> > > Page pool config is default, I don't know enough to fiddle yet.
> Also,
> > > the test is on a single stand alone server with no clustering, we
> are
> > > using Tomcat clustering for production.
> >
> > You should take a look a them, as they can affect the memory
> > consumption
> > directly.
> >
> > > I did fire the GC manually before getting concerned BTW. A lot of
> > heap
> > > seems to be taken up with 100-300k descriptions on the
> > > InternalClassTransformationImpl class. An example of one is
> attached
> > > below - seems wrong to me and more like the debug that can be
> > requested
> > > through log4j rather than a description.
> >
> > They look really bigger than they should. 300k is a lot. I have one
> > application with a reasonable number of pages and components and it
> > runs
> > inside a 512 MB slide at SliceHost without using the whole memory.
> >
> > --
> > Thiago H. de Paula Figueiredo
> > Independent Java, Apache Tapestry 5 and Hibernate consultant,
> > developer,
> > and instructor
> > Owner, software architect and developer, Ars Machina Tecnologia da
> > Informação Ltda.
> > http://www.arsmachina.com.br
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
> > For additional commands, e-mail: users-help@tapestry.apache.org
> >


RE: Tapestry using 1.3Gb of heap space after capacity testing

Posted by "Blower, Andy" <An...@proquest.co.uk>.
Thanks for the link. As I said before, I don't think that we've fallen prey to that exactly. Doing some local testing and looking at one of our largest pages which displays search results, it has 76 component objects, each of those will have components nested within (pretty heavily in some cases) but nothing which has a bunch of components where only one is used and the others are redundant. (what I think is this uber-component thing)

It looks like explicit repeated use of a component, rather than using a single component in a loop, creates a lot more component objects in the heap. I've never seen anything warning against this, and we do usually use loops but it's not always the best solution.

Again, looking at our results page, a single instance seems to get 11,283 ComponentPageElementImpl instances created for it. I'm finding it hard to find a good view of the tree to figure out what component(s) are the main culprits for this huge number of instances.

Thanks for any help or guidance you can give me.

Andy

> -----Original Message-----
> From: Thiago H. de Paula Figueiredo [mailto:thiagohp@gmail.com]
> Sent: 20 April 2010 15:56
> To: Tapestry users
> Subject: Re: Tapestry using 1.3Gb of heap space after capacity testing
> 
> On Tue, 20 Apr 2010 11:41:25 -0300, Blower, Andy
> <An...@proquest.co.uk> wrote:
> 
> > Uber-component anti-pattern, not to my knowledge, but we have a lot
> of
> > pages and components now. And a lot of logic in them. The project's
> > large enough I can't be sure. Do you have a handy link to Howard's
> > description so I can get my team to check?
> 
> It's here:
> http://old.nabble.com/-T5.0.18--Out-of-Memory-Error---Potential-Leak-
> (doesn't-reduce-after-forced-GC)-to25403474s302.html#a25497441
> That thread is interesting, by the way. The problem is more about how
> components and pages are written than the number of them.
> What Tapestry version are you using?
> 
> > Page pool config is default, I don't know enough to fiddle yet. Also,
> > the test is on a single stand alone server with no clustering, we are
> > using Tomcat clustering for production.
> 
> You should take a look a them, as they can affect the memory
> consumption
> directly.
> 
> > I did fire the GC manually before getting concerned BTW. A lot of
> heap
> > seems to be taken up with 100-300k descriptions on the
> > InternalClassTransformationImpl class. An example of one is attached
> > below - seems wrong to me and more like the debug that can be
> requested
> > through log4j rather than a description.
> 
> They look really bigger than they should. 300k is a lot. I have one
> application with a reasonable number of pages and components and it
> runs
> inside a 512 MB slide at SliceHost without using the whole memory.
> 
> --
> Thiago H. de Paula Figueiredo
> Independent Java, Apache Tapestry 5 and Hibernate consultant,
> developer,
> and instructor
> Owner, software architect and developer, Ars Machina Tecnologia da
> Informação Ltda.
> http://www.arsmachina.com.br
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
> For additional commands, e-mail: users-help@tapestry.apache.org
> 


Re: Tapestry using 1.3Gb of heap space after capacity testing

Posted by "Thiago H. de Paula Figueiredo" <th...@gmail.com>.
On Tue, 20 Apr 2010 11:41:25 -0300, Blower, Andy  
<An...@proquest.co.uk> wrote:

> Uber-component anti-pattern, not to my knowledge, but we have a lot of  
> pages and components now. And a lot of logic in them. The project's  
> large enough I can't be sure. Do you have a handy link to Howard's  
> description so I can get my team to check?

It's here:  
http://old.nabble.com/-T5.0.18--Out-of-Memory-Error---Potential-Leak-(doesn't-reduce-after-forced-GC)-to25403474s302.html#a25497441
That thread is interesting, by the way. The problem is more about how  
components and pages are written than the number of them.
What Tapestry version are you using?

> Page pool config is default, I don't know enough to fiddle yet. Also,  
> the test is on a single stand alone server with no clustering, we are  
> using Tomcat clustering for production.

You should take a look a them, as they can affect the memory consumption  
directly.

> I did fire the GC manually before getting concerned BTW. A lot of heap  
> seems to be taken up with 100-300k descriptions on the  
> InternalClassTransformationImpl class. An example of one is attached  
> below - seems wrong to me and more like the debug that can be requested  
> through log4j rather than a description.

They look really bigger than they should. 300k is a lot. I have one  
application with a reasonable number of pages and components and it runs  
inside a 512 MB slide at SliceHost without using the whole memory.

-- 
Thiago H. de Paula Figueiredo
Independent Java, Apache Tapestry 5 and Hibernate consultant, developer,  
and instructor
Owner, software architect and developer, Ars Machina Tecnologia da  
Informação Ltda.
http://www.arsmachina.com.br

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
For additional commands, e-mail: users-help@tapestry.apache.org


RE: Tapestry using 1.3Gb of heap space after capacity testing

Posted by "Blower, Andy" <An...@proquest.co.uk>.
> -----Original Message-----
> From: Thiago H. de Paula Figueiredo [mailto:thiagohp@gmail.com]
> Sent: 20 April 2010 15:30
> To: Tapestry users
> Subject: Re: Tapestry using 1.3Gb of heap space after capacity testing
>
> On Tue, 20 Apr 2010 11:16:11 -0300, Blower, Andy
> <An...@proquest.co.uk> wrote:
>
> > We've been doing some capacity testing recently, and I'm concerned
> about
> > the amount of heap memory that doesn't get freed when all the
> sessions
> > expire. I'm seeing 1.3Gb after being left with no use overnight.
>
> It seems you have huge pages and/or component. By any chance are you
> using
> the ûber-component antipattern described by Howard?
>
> What are your page pool configuration values?
>
> There was a message in this list about a Tapestry load test and it
> showed
> that the memory comsumption went back to normal after a large load.
>

Uber-component anti-pattern, not to my knowledge, but we have a lot of pages and components now. And a lot of logic in them. The project's large enough I can't be sure. Do you have a handy link to Howard's description so I can get my team to check?

Page pool config is default, I don't know enough to fiddle yet. Also, the test is on a single stand alone server with no clustering, we are using Tomcat clustering for production.

I did fire the GC manually before getting concerned BTW. A lot of heap seems to be taken up with 100-300k descriptions on the InternalClassTransformationImpl class. An example of one is attached below - seems wrong to me and more like the debug that can be requested through log4j rather than a description.


--------------------------------------------------------

add default method: public void postRenderCleanup()
<default>

add default method: public void setupRender(org.apache.tapestry5.MarkupWriter $1, org.apache.tapestry5.runtime.Event $2)
<default>

add default method: public void beginRender(org.apache.tapestry5.MarkupWriter $1, org.apache.tapestry5.runtime.Event $2)
<default>

add default method: public void beforeRenderTemplate(org.apache.tapestry5.MarkupWriter $1, org.apache.tapestry5.runtime.Event $2)
<default>

add default method: public void afterRenderTemplate(org.apache.tapestry5.MarkupWriter $1, org.apache.tapestry5.runtime.Event $2)
<default>

add default method: public void beforeRenderBody(org.apache.tapestry5.MarkupWriter $1, org.apache.tapestry5.runtime.Event $2)
<default>

add default method: public void afterRenderBody(org.apache.tapestry5.MarkupWriter $1, org.apache.tapestry5.runtime.Event $2)
<default>

add default method: public void afterRender(org.apache.tapestry5.MarkupWriter $1, org.apache.tapestry5.runtime.Event $2)
<default>

add default method: public void cleanupRender(org.apache.tapestry5.MarkupWriter $1, org.apache.tapestry5.runtime.Event $2)
<default>

add default method: public boolean dispatchComponentEvent(org.apache.tapestry5.runtime.ComponentEvent $1)
<default>

add default method: public org.apache.tapestry5.ComponentResources getComponentResources()
<default>

add default method: public void containingPageDidLoad()
<default>

add default method: public void containingPageDidDetach()
<default>

add default method: public void containingPageDidAttach()
<default>

add default method: public void restoreStateBeforePageAttach()
<default>

add field: protected final org.apache.tapestry5.internal.InternalComponentResources _$resources;

replace method: public final org.apache.tapestry5.ComponentResources getComponentResources()
return _$resources;

add method: private void _$write_LOGGER(org.slf4j.Logger $1)
throw new java.lang.RuntimeException("Field com.proquest.apps.corelib.components.MyResearchCheckbox.LOGGER is read-only.");

replace write LOGGER: _$write_LOGGER();

add method: private void _$write_durableUrlService(com.proquest.apps.corelib.util.DurableURL $1)
throw new java.lang.RuntimeException("Field com.proquest.apps.corelib.components.MyResearchCheckbox.durableUrlService is read-only.");

replace write durableUrlService: _$write_durableUrlService();

add method: private void _$write_isoLangService(com.proquest.apps.corelib.util.ISOLanguageService $1)
throw new java.lang.RuntimeException("Field com.proquest.apps.corelib.components.MyResearchCheckbox.isoLangService is read-only.");

replace write isoLangService: _$write_isoLangService();

add method: private void _$write_messages(org.apache.tapestry5.ioc.Messages $1)
throw new java.lang.RuntimeException("Field com.proquest.apps.corelib.components.MyResearchCheckbox.messages is read-only.");

replace write messages: _$write_messages();

add method: private void _$write_request(org.apache.tapestry5.services.Request $1)
throw new java.lang.RuntimeException("Field com.proquest.apps.corelib.components.MyResearchCheckbox.request is read-only.");

replace write request: _$write_request();

add method: private void _$write_resources(org.apache.tapestry5.ComponentResources $1)
throw new java.lang.RuntimeException("Field com.proquest.apps.corelib.components.MyResearchCheckbox.resources is read-only.");

replace write resources: _$write_resources();

add method: private void _$write_createFolderZone(org.apache.tapestry5.corelib.components.Zone $1)
throw new java.lang.RuntimeException("Field com.proquest.apps.corelib.components.MyResearchCheckbox.createFolderZone is read-only.");

replace write createFolderZone: _$write_createFolderZone();

extend method: public void containingPageDidLoad()
createFolderZone = (org.apache.tapestry5.corelib.components.Zone) _$resources.getEmbeddedComponent("createFolderZone");

extend method: public boolean dispatchComponentEvent(org.apache.tapestry5.runtime.ComponentEvent $1)
{
  if ($1.isAborted()) return $_;
  try
  {
    if ($1.matches("markItem", "", 1))
    {
      $_ = true;
      $1.setMethodDescription("com.proquest.apps.corelib.components.MyResearchCheckbox.mark(boolean) (at MyResearchCheckbox.java:181)");
      if ($1.storeResult(($w) mark(((java.lang.Boolean)$1.coerceContext(0, "java.lang.Boolean")).booleanValue()))) return true;
    }
    if ($1.matches("markAllItems", "", 1))
    {
      $_ = true;
      $1.setMethodDescription("com.proquest.apps.corelib.components.MyResearchCheckbox.markAll(boolean) (at MyResearchCheckbox.java:204)");
      if ($1.storeResult(($w) markAll(((java.lang.Boolean)$1.coerceContext(0, "java.lang.Boolean")).booleanValue()))) return true;
    }
  }
  catch (RuntimeException ex) { throw ex; }
  catch (Exception ex) { throw new RuntimeException(ex); }
}


add default method: public void render(org.apache.tapestry5.MarkupWriter $1, org.apache.tapestry5.runtime.RenderQueue $2)
<default>

replace method: public void render(org.apache.tapestry5.MarkupWriter $1, org.apache.tapestry5.runtime.RenderQueue $2)
_$resources.queueRender($2);

add field: private boolean _$attachedId_cached;

add field: private org.apache.tapestry5.internal.ParameterAccess _$attachedId_access;

add field: private java.lang.String _$attachedId_default;

extend method: public void containingPageDidLoad()
{
  _$attachedId_access = _$resources.getParameterAccess("attachedId");
  _$attachedId_default = attachedId;
}


extend method: public void postRenderCleanup()
if (! _$attachedId_access.isInvariant())
{
  attachedId = _$attachedId_default;
  _$attachedId_cached = false;
}


extend method: public void containingPageDidDetach()
if (! _$attachedId_access.isInvariant())
{
  attachedId = _$attachedId_default;
  _$attachedId_cached = false;
}


add method: private java.lang.String _$read_parameter_attachedId()
{
  if (_$attachedId_cached || ! _$resources.isLoaded() || ! _$attachedId_access.isBound()) return attachedId;
  java.lang.String result = ($r) ((java.lang.String) _$attachedId_access.read("java.lang.String"));
  if (_$attachedId_access.isInvariant() || _$resources.isRendering())
  {
    attachedId = result;
    _$attachedId_cached = true;
  }
  return result;
}


replace read attachedId: _$read_parameter_attachedId();

add method: private void _$update_parameter_attachedId(java.lang.String $1)
{
  if (! _$resources.isLoaded())
  {
    attachedId = $1;
    return;
  }
  _$attachedId_access.write(($w)$1);
  attachedId = $1;
  _$attachedId_cached = _$resources.isRendering();
}


replace write attachedId: _$update_parameter_attachedId();

add field: private boolean _$clientId_cached;

add field: private org.apache.tapestry5.internal.ParameterAccess _$clientId_access;

add field: private java.lang.String _$clientId_default;

extend method: public void containingPageDidLoad()
{
  _$clientId_access = _$resources.getParameterAccess("id");
  _$clientId_default = clientId;
}


extend method: public void postRenderCleanup()
if (! _$clientId_access.isInvariant())
{
  clientId = _$clientId_default;
  _$clientId_cached = false;
}


extend method: public void containingPageDidDetach()
if (! _$clientId_access.isInvariant())
{
  clientId = _$clientId_default;
  _$clientId_cached = false;
}


add method: private java.lang.String _$read_parameter_id()
{
  if (_$clientId_cached || ! _$resources.isLoaded() || ! _$clientId_access.isBound()) return clientId;
  java.lang.String result = ($r) ((java.lang.String) _$clientId_access.read("java.lang.String"));
  if (_$clientId_access.isInvariant() || _$resources.isRendering())
  {
    clientId = result;
    _$clientId_cached = true;
  }
  return result;
}


replace read clientId: _$read_parameter_id();

add method: private void _$update_parameter_id(java.lang.String $1)
{
  if (! _$resources.isLoaded())
  {
    clientId = $1;
    return;
  }
  _$clientId_access.write(($w)$1);
  clientId = $1;
  _$clientId_cached = _$resources.isRendering();
}


replace write clientId: _$update_parameter_id();

add field: private boolean _$displayLanguageIso3_cached;

add field: private org.apache.tapestry5.internal.ParameterAccess _$displayLanguageIso3_access;

add field: private java.lang.String _$displayLanguageIso3_default;

extend method: public void containingPageDidLoad()
{
  _$displayLanguageIso3_access = _$resources.getParameterAccess("displayLanguageIso3");
  _$displayLanguageIso3_default = displayLanguageIso3;
}


extend method: public void postRenderCleanup()
if (! _$displayLanguageIso3_access.isInvariant())
{
  displayLanguageIso3 = _$displayLanguageIso3_default;
  _$displayLanguageIso3_cached = false;
}


extend method: public void containingPageDidDetach()
if (! _$displayLanguageIso3_access.isInvariant())
{
  displayLanguageIso3 = _$displayLanguageIso3_default;
  _$displayLanguageIso3_cached = false;
}


add method: private java.lang.String _$read_parameter_displayLanguageIso3()
{
  if (_$displayLanguageIso3_cached || ! _$resources.isLoaded() || ! _$displayLanguageIso3_access.isBound()) return displayLanguageIso3;
  java.lang.String result = ($r) ((java.lang.String) _$displayLanguageIso3_access.read("java.lang.String"));
  if (_$displayLanguageIso3_access.isInvariant() || _$resources.isRendering())
  {
    displayLanguageIso3 = result;
    _$displayLanguageIso3_cached = true;
  }
  return result;
}


replace read displayLanguageIso3: _$read_parameter_displayLanguageIso3();

add method: private void _$update_parameter_displayLanguageIso3(java.lang.String $1)
{
  if (! _$resources.isLoaded())
  {
    displayLanguageIso3 = $1;
    return;
  }
  _$displayLanguageIso3_access.write(($w)$1);
  displayLanguageIso3 = $1;
  _$displayLanguageIso3_cached = _$resources.isRendering();
}


replace write displayLanguageIso3: _$update_parameter_displayLanguageIso3();

add field: private boolean _$itemId_cached;

add field: private org.apache.tapestry5.internal.ParameterAccess _$itemId_access;

add field: private com.proquest.apps.corelib.search.UniqueItemID _$itemId_default;

extend method: public void containingPageDidLoad()
{
  _$itemId_access = _$resources.getParameterAccess("itemId");
  _$itemId_default = itemId;
}


extend method: public void postRenderCleanup()
if (! _$itemId_access.isInvariant())
{
  itemId = _$itemId_default;
  _$itemId_cached = false;
}


extend method: public void containingPageDidDetach()
if (! _$itemId_access.isInvariant())
{
  itemId = _$itemId_default;
  _$itemId_cached = false;
}


add method: private com.proquest.apps.corelib.search.UniqueItemID _$read_parameter_itemId()
{
  if (_$itemId_cached || ! _$resources.isLoaded() || ! _$itemId_access.isBound()) return itemId;
  com.proquest.apps.corelib.search.UniqueItemID result = ($r) ((com.proquest.apps.corelib.search.UniqueItemID) _$itemId_access.read("com.proquest.apps.corelib.search.UniqueItemID"));
  if (_$itemId_access.isInvariant() || _$resources.isRendering())
  {
    itemId = result;
    _$itemId_cached = true;
  }
  return result;
}


replace read itemId: _$read_parameter_itemId();

add method: private void _$update_parameter_itemId(com.proquest.apps.corelib.search.UniqueItemID $1)
{
  if (! _$resources.isLoaded())
  {
    itemId = $1;
    return;
  }
  _$itemId_access.write(($w)$1);
  itemId = $1;
  _$itemId_cached = _$resources.isRendering();
}


replace write itemId: _$update_parameter_itemId();

add field: private boolean _$marked_cached;

add field: private org.apache.tapestry5.internal.ParameterAccess _$marked_access;

add field: private boolean _$marked_default;

add field: protected final org.apache.tapestry5.services.BindingSource _$bindingSource;

extend method: public void containingPageDidLoad()
{
  if (! _$resources.isBound("marked"))
    _$resources.bindParameter("marked", _$bindingSource.newBinding("default marked", _$resources, "prop", "false"));
  _$marked_access = _$resources.getParameterAccess("marked");
  _$marked_default = marked;
}


extend method: public void postRenderCleanup()
if (! _$marked_access.isInvariant())
{
  marked = _$marked_default;
  _$marked_cached = false;
}


extend method: public void containingPageDidDetach()
if (! _$marked_access.isInvariant())
{
  marked = _$marked_default;
  _$marked_cached = false;
}


add method: private boolean _$read_parameter_marked()
{
  if (_$marked_cached || ! _$resources.isLoaded() || ! _$marked_access.isBound()) return marked;
  boolean result = ($r) ((java.lang.Boolean) _$marked_access.read("java.lang.Boolean"));
  if (_$marked_access.isInvariant() || _$resources.isRendering())
  {
    marked = result;
    _$marked_cached = true;
  }
  return result;
}


replace read marked: _$read_parameter_marked();

add method: private void _$update_parameter_marked(boolean $1)
{
  if (! _$resources.isLoaded())
  {
    marked = $1;
    return;
  }
  _$marked_access.write(($w)$1);
  marked = $1;
  _$marked_cached = _$resources.isRendering();
}


replace write marked: _$update_parameter_marked();

add field: private boolean _$resultsPagedList_cached;

add field: private org.apache.tapestry5.internal.ParameterAccess _$resultsPagedList_access;

add field: private com.proquest.apps.corelib.components.ResultsPagedList _$resultsPagedList_default;

extend method: public void containingPageDidLoad()
{
  _$resultsPagedList_access = _$resources.getParameterAccess("for");
  _$resultsPagedList_default = resultsPagedList;
}


extend method: public void postRenderCleanup()
if (! _$resultsPagedList_access.isInvariant())
{
  resultsPagedList = _$resultsPagedList_default;
  _$resultsPagedList_cached = false;
}


extend method: public void containingPageDidDetach()
if (! _$resultsPagedList_access.isInvariant())
{
  resultsPagedList = _$resultsPagedList_default;
  _$resultsPagedList_cached = false;
}


add method: private com.proquest.apps.corelib.components.ResultsPagedList _$read_parameter_for()
{
  if (_$resultsPagedList_cached || ! _$resources.isLoaded() || ! _$resultsPagedList_access.isBound()) return resultsPagedList;
  com.proquest.apps.corelib.components.ResultsPagedList result = ($r) ((com.proquest.apps.corelib.components.ResultsPagedList) _$resultsPagedList_access.read("com.proquest.apps.corelib.components.ResultsPagedList"));
  if (_$resultsPagedList_access.isInvariant() || _$resources.isRendering())
  {
    resultsPagedList = result;
    _$resultsPagedList_cached = true;
  }
  return result;
}


replace read resultsPagedList: _$read_parameter_for();

add method: private void _$update_parameter_for(com.proquest.apps.corelib.components.ResultsPagedList $1)
{
  if (! _$resources.isLoaded())
  {
    resultsPagedList = $1;
    return;
  }
  _$resultsPagedList_access.write(($w)$1);
  resultsPagedList = $1;
  _$resultsPagedList_cached = _$resources.isRendering();
}


replace write resultsPagedList: _$update_parameter_for();

add field: private boolean _$selectDestinationArea_cached;

add field: private org.apache.tapestry5.internal.ParameterAccess _$selectDestinationArea_access;

add field: private boolean _$selectDestinationArea_default;

extend method: public void containingPageDidLoad()
{
  if (! _$resources.isBound("selectDestinationArea"))
    _$resources.bindParameter("selectDestinationArea", _$bindingSource.newBinding("default selectDestinationArea", _$resources, "prop", "false"));
  _$selectDestinationArea_access = _$resources.getParameterAccess("selectDestinationArea");
  _$selectDestinationArea_default = selectDestinationArea;
}


extend method: public void postRenderCleanup()
if (! _$selectDestinationArea_access.isInvariant())
{
  selectDestinationArea = _$selectDestinationArea_default;
  _$selectDestinationArea_cached = false;
}


extend method: public void containingPageDidDetach()
if (! _$selectDestinationArea_access.isInvariant())
{
  selectDestinationArea = _$selectDestinationArea_default;
  _$selectDestinationArea_cached = false;
}


add method: private boolean _$read_parameter_selectDestinationArea()
{
  if (_$selectDestinationArea_cached || ! _$resources.isLoaded() || ! _$selectDestinationArea_access.isBound()) return selectDestinationArea;
  boolean result = ($r) ((java.lang.Boolean) _$selectDestinationArea_access.read("java.lang.Boolean"));
  if (_$selectDestinationArea_access.isInvariant() || _$resources.isRendering())
  {
    selectDestinationArea = result;
    _$selectDestinationArea_cached = true;
  }
  return result;
}


replace read selectDestinationArea: _$read_parameter_selectDestinationArea();

add method: private void _$update_parameter_selectDestinationArea(boolean $1)
{
  if (! _$resources.isLoaded())
  {
    selectDestinationArea = $1;
    return;
  }
  _$selectDestinationArea_access.write(($w)$1);
  selectDestinationArea = $1;
  _$selectDestinationArea_cached = _$resources.isRendering();
}


replace write selectDestinationArea: _$update_parameter_selectDestinationArea();

replace method: public void beginRender(org.apache.tapestry5.MarkupWriter $1, org.apache.tapestry5.runtime.Event $2)
{
  try
  {
    $2.setMethodDescription("com.proquest.apps.corelib.components.MyResearchCheckbox.render(org.apache.tapestry5.MarkupWriter) (at MyResearchCheckbox.java:123)");
    if ($2.storeResult(($w) render($1))) return;
  }
  catch (RuntimeException ex) { throw ex; }
  catch (Exception ex) { throw new RuntimeException(ex); }
}


add field: protected final org.apache.tapestry5.internal.services.ComponentResourcesOperation _$operation;

extend method: public void setupRender(org.apache.tapestry5.MarkupWriter $1, org.apache.tapestry5.runtime.Event $2)
_$operation.perform(_$resources);

extend method: public void postRenderCleanup()
_$resources.postRenderCleanup();

add transformed method: public final java.lang.String getClientIdProperty()
return clientIdProperty;

add transformed method: public final void setClientIdProperty(java.lang.String $1)
clientIdProperty = $1;

add transformed method: public final org.apache.tapestry5.corelib.components.Zone getCreateFolderZone()
return createFolderZone;

add transformed method: public final void setCreateFolderZone(org.apache.tapestry5.corelib.components.Zone $1)
createFolderZone = $1;

add field: protected final org.apache.tapestry5.services.ApplicationStateManager _$applicationStateManager;

add field: protected final java.lang.Class _$myResearchItems_type;

add method: private com.proquest.apps.corelib.myresearch.MyResearchItems _$read_myResearchItems()
return (com.proquest.apps.corelib.myresearch.MyResearchItems) _$applicationStateManager.get(_$myResearchItems_type);

replace read myResearchItems: _$read_myResearchItems();

add method: private void _$write_myResearchItems(com.proquest.apps.corelib.myresearch.MyResearchItems $1)
_$applicationStateManager.set(_$myResearchItems_type, $1);

replace write myResearchItems: _$write_myResearchItems();

remove field myResearchItems;

add field: protected final org.apache.tapestry5.services.Environment _$environment;

add field: protected final java.lang.Class _$type;

add field: private org.apache.tapestry5.services.EnvironmentalAccess _$renderSupport_access;

extend method: public void containingPageDidAttach()
_$renderSupport_access = _$environment.getAccess(_$type);

extend method: public void containingPageDidDetach()
_$renderSupport_access = null;

add method: private org.apache.tapestry5.RenderSupport _$environment_read_renderSupport()
return ($r) _$renderSupport_access.peekRequired();

replace read renderSupport: _$environment_read_renderSupport();

add method: private void _$write_renderSupport(org.apache.tapestry5.RenderSupport $1)
throw new java.lang.RuntimeException("Field com.proquest.apps.corelib.components.MyResearchCheckbox.renderSupport is read-only.");

replace write renderSupport: _$write_renderSupport();

remove field renderSupport;

add field: private java.lang.String _$clientApp_default;

extend method: public void containingPageDidLoad()
_$clientApp_default = clientApp;

extend method: public void containingPageDidDetach()
clientApp = _$clientApp_default;

add field: private java.lang.String _$clientIdProperty_default;

extend method: public void containingPageDidLoad()
_$clientIdProperty_default = clientIdProperty;

extend method: public void containingPageDidDetach()
clientIdProperty = _$clientIdProperty_default;

add field: private java.lang.String _$database_default;

extend method: public void containingPageDidLoad()
_$database_default = database;

extend method: public void containingPageDidDetach()
database = _$database_default;

convert default constructor: initializer();

Checking field read clientId in method getClientId(): replacing with $_ = $0._$read_parameter_id();
Checking field read resources in method render(): field not transformed
Checking field read renderSupport in method render(): replacing with $_ = $0._$environment_read_renderSupport();
Checking field read resources in method render(): field not transformed
Checking field write clientId in method render(): replacing with $0._$update_parameter_id($1);
Checking field read clientId in method render(): replacing with $_ = $0._$read_parameter_id();
Checking field write clientIdProperty in method render(): field not transformed
Checking field read attachedId in method render(): replacing with $_ = $0._$read_parameter_attachedId();
Checking field read itemId in method render(): replacing with $_ = $0._$read_parameter_itemId();
Checking field read resultsPagedList in method render(): replacing with $_ = $0._$read_parameter_for();
Checking field read myResearchItems in method render(): replacing with $_ = $0._$read_myResearchItems();
Checking field read resources in method render(): field not transformed
Checking field read itemId in method render(): replacing with $_ = $0._$read_parameter_itemId();
Checking field read resources in method render(): field not transformed
Checking field read marked in method render(): replacing with $_ = $0._$read_parameter_marked();
Checking field read resources in method render(): field not transformed
Checking field read renderSupport in method render(): replacing with $_ = $0._$environment_read_renderSupport();
Checking field read request in method mark(): field not transformed
Checking field read request in method mark(): field not transformed
Checking field read request in method markAll(): field not transformed
Checking field read request in method markAll(): field not transformed
Checking field read myResearchItems in method markAll(): replacing with $_ = $0._$read_myResearchItems();
Checking field read myResearchItems in method markAll(): replacing with $_ = $0._$read_myResearchItems();
Checking field read myResearchItems in method markAll(): replacing with $_ = $0._$read_myResearchItems();
Checking field read LOGGER in method markItem(): field not transformed
Checking field read durableUrlService in method markItem(): field not transformed
Checking field read clientApp in method markItem(): field not transformed
Checking field read isoLangService in method markItem(): field not transformed
Checking field read displayLanguageIso3 in method markItem(): replacing with $_ = $0._$read_parameter_displayLanguageIso3();
Checking field read displayLanguageIso3 in method markItem(): replacing with $_ = $0._$read_parameter_displayLanguageIso3();
Checking field read AllFigTables in method markItem(): field not transformed
Checking field read myResearchItems in method markItem(): replacing with $_ = $0._$read_myResearchItems();
Checking field read AllFigTables in method markItem(): field not transformed
Checking field read All in method markItem(): field not transformed
Checking field read myResearchItems in method markItem(): replacing with $_ = $0._$read_myResearchItems();
Checking field read All in method markItem(): field not transformed
Checking field read LOGGER in method markItem(): field not transformed
Checking field read myResearchItems in method markItem(): replacing with $_ = $0._$read_myResearchItems();
Checking field read messages in method getNumInMyResearchMsg(): field not transformed
Checking field read myResearchItems in method getNumInMyResearchMsg(): replacing with $_ = $0._$read_myResearchItems();
Checking field read myResearchItems in method getNumInMyResearchMsg(): replacing with $_ = $0._$read_myResearchItems();
Checking field read clientApp in method getClientApp(): field not transformed
Checking field write clientApp in method setClientApp(): field not transformed
Checking field write database in method setDatabase(): field not transformed
Checking field read database in method getDatabase(): field not transformed
Checking field read _$attachedId_access in method postRenderCleanup(): added method
Checking field read _$attachedId_default in method postRenderCleanup(): added method
Checking field write attachedId in method postRenderCleanup(): added method
Checking field write _$attachedId_cached in method postRenderCleanup(): added method
Checking field read _$clientId_access in method postRenderCleanup(): added method
Checking field read _$clientId_default in method postRenderCleanup(): added method
Checking field write clientId in method postRenderCleanup(): added method
Checking field write _$clientId_cached in method postRenderCleanup(): added method
Checking field read _$displayLanguageIso3_access in method postRenderCleanup(): added method
Checking field read _$displayLanguageIso3_default in method postRenderCleanup(): added method
Checking field write displayLanguageIso3 in method postRenderCleanup(): added method
Checking field write _$displayLanguageIso3_cached in method postRenderCleanup(): added method
Checking field read _$itemId_access in method postRenderCleanup(): added method
Checking field read _$itemId_default in method postRenderCleanup(): added method
Checking field write itemId in method postRenderCleanup(): added method
Checking field write _$itemId_cached in method postRenderCleanup(): added method
Checking field read _$marked_access in method postRenderCleanup(): added method
Checking field read _$marked_default in method postRenderCleanup(): added method
Checking field write marked in method postRenderCleanup(): added method
Checking field write _$marked_cached in method postRenderCleanup(): added method
Checking field read _$resultsPagedList_access in method postRenderCleanup(): added method
Checking field read _$resultsPagedList_default in method postRenderCleanup(): added method
Checking field write resultsPagedList in method postRenderCleanup(): added method
Checking field write _$resultsPagedList_cached in method postRenderCleanup(): added method
Checking field read _$selectDestinationArea_access in method postRenderCleanup(): added method
Checking field read _$selectDestinationArea_default in method postRenderCleanup(): added method
Checking field write selectDestinationArea in method postRenderCleanup(): added method
Checking field write _$selectDestinationArea_cached in method postRenderCleanup(): added method
Checking field read _$resources in method postRenderCleanup(): added method
Checking field read _$operation in method setupRender(): added method
Checking field read _$resources in method setupRender(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field write createFolderZone in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field write _$attachedId_access in method containingPageDidLoad(): added method
Checking field read attachedId in method containingPageDidLoad(): added method
Checking field write _$attachedId_default in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field write _$clientId_access in method containingPageDidLoad(): added method
Checking field read clientId in method containingPageDidLoad(): added method
Checking field write _$clientId_default in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field write _$displayLanguageIso3_access in method containingPageDidLoad(): added method
Checking field read displayLanguageIso3 in method containingPageDidLoad(): added method
Checking field write _$displayLanguageIso3_default in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field write _$itemId_access in method containingPageDidLoad(): added method
Checking field read itemId in method containingPageDidLoad(): added method
Checking field write _$itemId_default in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field read _$bindingSource in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field write _$marked_access in method containingPageDidLoad(): added method
Checking field read marked in method containingPageDidLoad(): added method
Checking field write _$marked_default in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field write _$resultsPagedList_access in method containingPageDidLoad(): added method
Checking field read resultsPagedList in method containingPageDidLoad(): added method
Checking field write _$resultsPagedList_default in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field read _$bindingSource in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field read _$resources in method containingPageDidLoad(): added method
Checking field write _$selectDestinationArea_access in method containingPageDidLoad(): added method
Checking field read selectDestinationArea in method containingPageDidLoad(): added method
Checking field write _$selectDestinationArea_default in method containingPageDidLoad(): added method
Checking field read clientApp in method containingPageDidLoad(): added method
Checking field write _$clientApp_default in method containingPageDidLoad(): added method
Checking field read clientIdProperty in method containingPageDidLoad(): added method
Checking field write _$clientIdProperty_default in method containingPageDidLoad(): added method
Checking field read database in method containingPageDidLoad(): added method
Checking field write _$database_default in method containingPageDidLoad(): added method
Checking field read _$attachedId_access in method containingPageDidDetach(): added method
Checking field read _$attachedId_default in method containingPageDidDetach(): added method
Checking field write attachedId in method containingPageDidDetach(): added method
Checking field write _$attachedId_cached in method containingPageDidDetach(): added method
Checking field read _$clientId_access in method containingPageDidDetach(): added method
Checking field read _$clientId_default in method containingPageDidDetach(): added method
Checking field write clientId in method containingPageDidDetach(): added method
Checking field write _$clientId_cached in method containingPageDidDetach(): added method
Checking field read _$displayLanguageIso3_access in method containingPageDidDetach(): added method
Checking field read _$displayLanguageIso3_default in method containingPageDidDetach(): added method
Checking field write displayLanguageIso3 in method containingPageDidDetach(): added method
Checking field write _$displayLanguageIso3_cached in method containingPageDidDetach(): added method
Checking field read _$itemId_access in method containingPageDidDetach(): added method
Checking field read _$itemId_default in method containingPageDidDetach(): added method
Checking field write itemId in method containingPageDidDetach(): added method
Checking field write _$itemId_cached in method containingPageDidDetach(): added method
Checking field read _$marked_access in method containingPageDidDetach(): added method
Checking field read _$marked_default in method containingPageDidDetach(): added method
Checking field write marked in method containingPageDidDetach(): added method
Checking field write _$marked_cached in method containingPageDidDetach(): added method
Checking field read _$resultsPagedList_access in method containingPageDidDetach(): added method
Checking field read _$resultsPagedList_default in method containingPageDidDetach(): added method
Checking field write resultsPagedList in method containingPageDidDetach(): added method
Checking field write _$resultsPagedList_cached in method containingPageDidDetach(): added method
Checking field read _$selectDestinationArea_access in method containingPageDidDetach(): added method
Checking field read _$selectDestinationArea_default in method containingPageDidDetach(): added method
Checking field write selectDestinationArea in method containingPageDidDetach(): added method
Checking field write _$selectDestinationArea_cached in method containingPageDidDetach(): added method
Checking field write _$renderSupport_access in method containingPageDidDetach(): added method
Checking field read _$clientApp_default in method containingPageDidDetach(): added method
Checking field write clientApp in method containingPageDidDetach(): added method
Checking field read _$clientIdProperty_default in method containingPageDidDetach(): added method
Checking field write clientIdProperty in method containingPageDidDetach(): added method
Checking field read _$database_default in method containingPageDidDetach(): added method
Checking field write database in method containingPageDidDetach(): added method
Checking field read _$environment in method containingPageDidAttach(): added method
Checking field read _$type in method containingPageDidAttach(): added method
Checking field write _$renderSupport_access in method containingPageDidAttach(): added method
Checking field read _$resources in method getComponentResources(): added method
Checking field read _$resources in method render(): added method
Checking field read _$attachedId_cached in method _$read_parameter_attachedId(): added method
Checking field read _$resources in method _$read_parameter_attachedId(): added method
Checking field read _$attachedId_access in method _$read_parameter_attachedId(): added method
Checking field read attachedId in method _$read_parameter_attachedId(): added method
Checking field read _$attachedId_access in method _$read_parameter_attachedId(): added method
Checking field read _$attachedId_access in method _$read_parameter_attachedId(): added method
Checking field read _$resources in method _$read_parameter_attachedId(): added method
Checking field write attachedId in method _$read_parameter_attachedId(): added method
Checking field write _$attachedId_cached in method _$read_parameter_attachedId(): added method
Checking field read _$resources in method _$update_parameter_attachedId(): added method
Checking field write attachedId in method _$update_parameter_attachedId(): added method
Checking field read _$attachedId_access in method _$update_parameter_attachedId(): added method
Checking field write attachedId in method _$update_parameter_attachedId(): added method
Checking field read _$resources in method _$update_parameter_attachedId(): added method
Checking field write _$attachedId_cached in method _$update_parameter_attachedId(): added method
Checking field read _$clientId_cached in method _$read_parameter_id(): added method
Checking field read _$resources in method _$read_parameter_id(): added method
Checking field read _$clientId_access in method _$read_parameter_id(): added method
Checking field read clientId in method _$read_parameter_id(): added method
Checking field read _$clientId_access in method _$read_parameter_id(): added method
Checking field read _$clientId_access in method _$read_parameter_id(): added method
Checking field read _$resources in method _$read_parameter_id(): added method
Checking field write clientId in method _$read_parameter_id(): added method
Checking field write _$clientId_cached in method _$read_parameter_id(): added method
Checking field read _$resources in method _$update_parameter_id(): added method
Checking field write clientId in method _$update_parameter_id(): added method
Checking field read _$clientId_access in method _$update_parameter_id(): added method
Checking field write clientId in method _$update_parameter_id(): added method
Checking field read _$resources in method _$update_parameter_id(): added method
Checking field write _$clientId_cached in method _$update_parameter_id(): added method
Checking field read _$displayLanguageIso3_cached in method _$read_parameter_displayLanguageIso3(): added method
Checking field read _$resources in method _$read_parameter_displayLanguageIso3(): added method
Checking field read _$displayLanguageIso3_access in method _$read_parameter_displayLanguageIso3(): added method
Checking field read displayLanguageIso3 in method _$read_parameter_displayLanguageIso3(): added method
Checking field read _$displayLanguageIso3_access in method _$read_parameter_displayLanguageIso3(): added method
Checking field read _$displayLanguageIso3_access in method _$read_parameter_displayLanguageIso3(): added method
Checking field read _$resources in method _$read_parameter_displayLanguageIso3(): added method
Checking field write displayLanguageIso3 in method _$read_parameter_displayLanguageIso3(): added method
Checking field write _$displayLanguageIso3_cached in method _$read_parameter_displayLanguageIso3(): added method
Checking field read _$resources in method _$update_parameter_displayLanguageIso3(): added method
Checking field write displayLanguageIso3 in method _$update_parameter_displayLanguageIso3(): added method
Checking field read _$displayLanguageIso3_access in method _$update_parameter_displayLanguageIso3(): added method
Checking field write displayLanguageIso3 in method _$update_parameter_displayLanguageIso3(): added method
Checking field read _$resources in method _$update_parameter_displayLanguageIso3(): added method
Checking field write _$displayLanguageIso3_cached in method _$update_parameter_displayLanguageIso3(): added method
Checking field read _$itemId_cached in method _$read_parameter_itemId(): added method
Checking field read _$resources in method _$read_parameter_itemId(): added method
Checking field read _$itemId_access in method _$read_parameter_itemId(): added method
Checking field read itemId in method _$read_parameter_itemId(): added method
Checking field read _$itemId_access in method _$read_parameter_itemId(): added method
Checking field read _$itemId_access in method _$read_parameter_itemId(): added method
Checking field read _$resources in method _$read_parameter_itemId(): added method
Checking field write itemId in method _$read_parameter_itemId(): added method
Checking field write _$itemId_cached in method _$read_parameter_itemId(): added method
Checking field read _$resources in method _$update_parameter_itemId(): added method
Checking field write itemId in method _$update_parameter_itemId(): added method
Checking field read _$itemId_access in method _$update_parameter_itemId(): added method
Checking field write itemId in method _$update_parameter_itemId(): added method
Checking field read _$resources in method _$update_parameter_itemId(): added method
Checking field write _$itemId_cached in method _$update_parameter_itemId(): added method
Checking field read _$marked_cached in method _$read_parameter_marked(): added method
Checking field read _$resources in method _$read_parameter_marked(): added method
Checking field read _$marked_access in method _$read_parameter_marked(): added method
Checking field read marked in method _$read_parameter_marked(): added method
Checking field read _$marked_access in method _$read_parameter_marked(): added method
Checking field read _$marked_access in method _$read_parameter_marked(): added method
Checking field read _$resources in method _$read_parameter_marked(): added method
Checking field write marked in method _$read_parameter_marked(): added method
Checking field write _$marked_cached in method _$read_parameter_marked(): added method
Checking field read _$resources in method _$update_parameter_marked(): added method
Checking field write marked in method _$update_parameter_marked(): added method
Checking field read _$marked_access in method _$update_parameter_marked(): added method
Checking field write marked in method _$update_parameter_marked(): added method
Checking field read _$resources in method _$update_parameter_marked(): added method
Checking field write _$marked_cached in method _$update_parameter_marked(): added method
Checking field read _$resultsPagedList_cached in method _$read_parameter_for(): added method
Checking field read _$resources in method _$read_parameter_for(): added method
Checking field read _$resultsPagedList_access in method _$read_parameter_for(): added method
Checking field read resultsPagedList in method _$read_parameter_for(): added method
Checking field read _$resultsPagedList_access in method _$read_parameter_for(): added method
Checking field read _$resultsPagedList_access in method _$read_parameter_for(): added method
Checking field read _$resources in method _$read_parameter_for(): added method
Checking field write resultsPagedList in method _$read_parameter_for(): added method
Checking field write _$resultsPagedList_cached in method _$read_parameter_for(): added method
Checking field read _$resources in method _$update_parameter_for(): added method
Checking field write resultsPagedList in method _$update_parameter_for(): added method
Checking field read _$resultsPagedList_access in method _$update_parameter_for(): added method
Checking field write resultsPagedList in method _$update_parameter_for(): added method
Checking field read _$resources in method _$update_parameter_for(): added method
Checking field write _$resultsPagedList_cached in method _$update_parameter_for(): added method
Checking field read _$selectDestinationArea_cached in method _$read_parameter_selectDestinationArea(): added method
Checking field read _$resources in method _$read_parameter_selectDestinationArea(): added method
Checking field read _$selectDestinationArea_access in method _$read_parameter_selectDestinationArea(): added method
Checking field read selectDestinationArea in method _$read_parameter_selectDestinationArea(): added method
Checking field read _$selectDestinationArea_access in method _$read_parameter_selectDestinationArea(): added method
Checking field read _$selectDestinationArea_access in method _$read_parameter_selectDestinationArea(): added method
Checking field read _$resources in method _$read_parameter_selectDestinationArea(): added method
Checking field write selectDestinationArea in method _$read_parameter_selectDestinationArea(): added method
Checking field write _$selectDestinationArea_cached in method _$read_parameter_selectDestinationArea(): added method
Checking field read _$resources in method _$update_parameter_selectDestinationArea(): added method
Checking field write selectDestinationArea in method _$update_parameter_selectDestinationArea(): added method
Checking field read _$selectDestinationArea_access in method _$update_parameter_selectDestinationArea(): added method
Checking field write selectDestinationArea in method _$update_parameter_selectDestinationArea(): added method
Checking field read _$resources in method _$update_parameter_selectDestinationArea(): added method
Checking field write _$selectDestinationArea_cached in method _$update_parameter_selectDestinationArea(): added method
Checking field read clientIdProperty in method getClientIdProperty(): field not transformed
Checking field write clientIdProperty in method setClientIdProperty(): field not transformed
Checking field read createFolderZone in method getCreateFolderZone(): field not transformed
Checking field write createFolderZone in method setCreateFolderZone(): replacing with $0._$write_createFolderZone($1);
Checking field read _$applicationStateManager in method _$read_myResearchItems(): added method
Checking field read _$myResearchItems_type in method _$read_myResearchItems(): added method
Checking field read _$applicationStateManager in method _$write_myResearchItems(): added method
Checking field read _$myResearchItems_type in method _$write_myResearchItems(): added method
Checking field read _$renderSupport_access in method _$environment_read_renderSupport(): added method

add constructor: com.proquest.apps.corelib.components.MyResearchCheckbox(org.apache.tapestry5.internal.InternalComponentResources $1, com.proquest.apps.corelib.util.DurableURL $2, com.proquest.apps.corelib.util.ISOLanguageService $3, org.apache.tapestry5.services.Request $4, org.apache.tapestry5.services.BindingSource $5, org.apache.tapestry5.internal.services.ComponentResourcesOperation $6, org.apache.tapestry5.services.ApplicationStateManager $7, java.lang.Class $8, org.apache.tapestry5.services.Environment $9, java.lang.Class $10)
{
  _$resources = $1;
LOGGER = _$resources.getLogger();
  durableUrlService = $2;
  isoLangService = $3;
messages = _$resources.getMessages();
  request = $4;
resources = _$resources;
  _$bindingSource = $5;
  _$operation = $6;
  _$applicationStateManager = $7;
  _$myResearchItems_type = $8;
  _$environment = $9;
  _$type = $10;
  initializer();

}



Re: Tapestry using 1.3Gb of heap space after capacity testing

Posted by "Thiago H. de Paula Figueiredo" <th...@gmail.com>.
On Tue, 20 Apr 2010 11:16:11 -0300, Blower, Andy  
<An...@proquest.co.uk> wrote:

> We've been doing some capacity testing recently, and I'm concerned about  
> the amount of heap memory that doesn't get freed when all the sessions  
> expire. I'm seeing 1.3Gb after being left with no use overnight.

It seems you have huge pages and/or component. By any chance are you using  
the ûber-component antipattern described by Howard?

What are your page pool configuration values?

There was a message in this list about a Tapestry load test and it showed  
that the memory comsumption went back to normal after a large load.

-- 
Thiago H. de Paula Figueiredo
Independent Java, Apache Tapestry 5 and Hibernate consultant, developer,  
and instructor
Owner, software architect and developer, Ars Machina Tecnologia da  
Informação Ltda.
http://www.arsmachina.com.br

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tapestry.apache.org
For additional commands, e-mail: users-help@tapestry.apache.org