You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lenya.apache.org by bu...@apache.org on 2007/08/29 18:05:11 UTC

DO NOT REPLY [Bug 43240] New: - Memory leak

DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240

           Summary: Memory leak
           Product: Lenya
           Version: Trunk
          Platform: Other
        OS/Version: other
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Miscellaneous
        AssignedTo: dev@lenya.apache.org
        ReportedBy: andreas@apache.org


It is very likely that we have a memory leak in the trunk.

Load test:
- Tomcat: Xms64m, Xmx512m
- 10 threads
- each creating 50 documents

The heap usage is continuously increasing. After approx. 800 requests, 400 MB of
heap are allocated. The garbage collector can't push it below 300 MB anymore. A
first glance at the heap snapshot indicates a very large number of
TraxTransformer instances.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 13:54 -------
org.apache.lenya.cms.repository.SessionImpl
session of  org.apache.lenya.cms.publication.DocumentFactoryImpl
factory of  org.apache.lenya.cms.publication.PublicationImpl
publication of  org.apache.lenya.cms.publication.DocumentIdentifier
key of  java.util.HashMap$Entry
[254] of  java.util.HashMap$Entry[257]
table of  java.util.HashMap
identifier2listeners of org.apache.lenya.cms.observation.ObservationManager

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 15:51 -------
(In reply to comment #13)
> (In reply to comment #12)
> > org.apache.lenya.cms.repository.SessionImpl
> > session of  org.apache.lenya.cms.publication.DocumentFactoryImpl
> > factory of  org.apache.lenya.cms.publication.PublicationImpl
> > publication of  org.apache.lenya.cms.publication.DocumentIdentifier
> 
> DocumentIdentifier shouldn't keep a reference to the publication, but only to
> the publication ID.

This is changed in revision 570957.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 14:03 -------
(In reply to comment #12)
> org.apache.lenya.cms.repository.SessionImpl
> session of  org.apache.lenya.cms.publication.DocumentFactoryImpl
> factory of  org.apache.lenya.cms.publication.PublicationImpl
> publication of  org.apache.lenya.cms.publication.DocumentIdentifier

DocumentIdentifier shouldn't keep a reference to the publication, but only to
the publication ID.

> key of  java.util.HashMap$Entry
> [254] of  java.util.HashMap$Entry[257]
> table of  java.util.HashMap
> identifier2listeners of org.apache.lenya.cms.observation.ObservationManager



-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


Re: DO NOT REPLY [Bug 43240] - Memory leak

Posted by Richard Frovarp <Ri...@sendit.nodak.edu>.
bugzilla@apache.org wrote:
>
> http://issues.apache.org/bugzilla/show_bug.cgi?id=43240
>
>
>
>
>
> ------- Additional Comments From andreas@apache.org  2007-08-30 04:19 -------
> I just ran the test with 20 threads, each creating 50 documents. It finished
> without an OOME.
>
>   
Thanks for working on this.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-30 04:19 -------
I just ran the test with 20 threads, each creating 50 documents. It finished
without an OOME.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 09:43 -------
When Tomcat doesn't receive any requests for some time (I guess because JMeter
is swapped to the disk due to RAM shortage), the heap usage decreases to 254 MB.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 13:37 -------
Now the largest object is the ContinuationsManagerImpl with a retained size of
34 MB, followed by the DefaultTransientStore with 17 MB.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 09:44 -------
After the 1020 requests are completed, the heap usage stays at approx. 300 MB.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240


andreas@apache.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |blocker
   Target Milestone|2.0.1                       |2.0




-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 13:51 -------
There are still more than 800 SessionImpl objects, connected via publication
objects as listeners of the ObservationManager.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 16:10 -------
Now there are 444 SessionImpl objects, attached via the continuation:

org.apache.lenya.cms.repository.SessionImpl
value of  java.util.HashMap$Entry
[57] of  java.util.HashMap$Entry[65]
table of  java.util.HashMap
parameters of  org.apache.lenya.cms.usecase.impl.UsecaseProxy
javaObject of  org.mozilla.javascript.NativeJavaObject
[1] of  java.lang.Object[9]
stack of  org.mozilla.javascript.Interpreter$CallFrame
implementation of  org.mozilla.javascript.continuations.Continuation
continuation of  org.apache.cocoon.components.flow.WebContinuation

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 16:21 -------
But why did 444 sessions create 82.456 DocumentImpl objects? I can't imagine
that all of them have been requested by the Create usecase. Does the sitetree
acquire DocumentImpl objects unnecessarily?

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 16:17 -------
After another half hour, a GC pushes the heap usage down to 184 MB. I guess some
of the continuations have been removed.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 13:28 -------
BTW, I increased the xslt transformer pool size to 64 which makes the test run a
lot faster, but will certainly influence the memory usage. I'll keep this
setting for further test.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


Re: When to use createContinuation="false" (was Re: [Bug 43240] Memory leak)

Posted by Andreas Hartmann <an...@apache.org>.
Thorsten Scherler schrieb:
> On Tue, 2008-02-26 at 16:02 -0800, bugzilla@apache.org wrote:
>> http://issues.apache.org/bugzilla/show_bug.cgi?id=43240
> 
>> I did some memory profiling with the MultiWorkflow usecase and noticed two
>> problems:
>>
>> 1. Continuations aren't invalidated and eat up memory.
>> 2. The MultiWorkflow usecase creates a continuation but doesn't need one.
>>
>> I have added some code to invalidate the continuation to usecases.js and added
>> the createContinuation="false" attribute to the MultiWorkflow view declaration.
>> We should be very careful to add this attribute to all view declarations which
>> don't need a continuation.
> 
> Is there a general rule of thumb when a usecase needs continuation and
> when not? I guess having this will lower the impact.

A usecase view needs a continuation if the usecase reacts to user 
interaction. This is usually the case if it implements advance() and/or 
doExecute(). I guess we should add this to the documentation.

-- Andreas



-- 
Andreas Hartmann, CTO
BeCompany GmbH
http://www.becompany.ch
Tel.: +41 (0) 43 818 57 01


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


When to use createContinuation="false" (was Re: [Bug 43240] Memory leak)

Posted by Thorsten Scherler <th...@juntadeandalucia.es>.
On Tue, 2008-02-26 at 16:02 -0800, bugzilla@apache.org wrote:
> http://issues.apache.org/bugzilla/show_bug.cgi?id=43240

> I did some memory profiling with the MultiWorkflow usecase and noticed two
> problems:
> 
> 1. Continuations aren't invalidated and eat up memory.
> 2. The MultiWorkflow usecase creates a continuation but doesn't need one.
> 
> I have added some code to invalidate the continuation to usecases.js and added
> the createContinuation="false" attribute to the MultiWorkflow view declaration.
> We should be very careful to add this attribute to all view declarations which
> don't need a continuation.

Is there a general rule of thumb when a usecase needs continuation and
when not? I guess having this will lower the impact.

salu2
-- 
Thorsten Scherler                                 thorsten.at.apache.org
Open Source Java                      consulting, training and solutions


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


[Bug 43240] Memory leak

Posted by bu...@apache.org.
http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





--- Comment #24 from Andreas Hartmann <an...@apache.org>  2008-02-26 16:02:30 ---
I did some memory profiling with the MultiWorkflow usecase and noticed two
problems:

1. Continuations aren't invalidated and eat up memory.
2. The MultiWorkflow usecase creates a continuation but doesn't need one.

I have added some code to invalidate the continuation to usecases.js and added
the createContinuation="false" attribute to the MultiWorkflow view declaration.
We should be very careful to add this attribute to all view declarations which
don't need a continuation.


-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 13:25 -------
Now the heap usage after the forced garbage collection (after running the test)
is at 138 MB.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240


rfrovarp@apache.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|blocker                     |critical
   Target Milestone|2.0                         |2.0.1




------- Additional Comments From rfrovarp@apache.org  2007-12-10 07:40 -------
I've been watching usage via jconsole. Perm gen sites at 52 MB or 65 MB
allocated. We may want to increase the permgen value to something a little
bigger. I need to do further testing, but exceptions seem to eat up massive
amounts of memory. I've been running for a week and "PS Old Gen" is pretty
stable at between 200 and 250 MB.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240


andreas@apache.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
OtherBugsDependingO|                            |43160
              nThis|                            |




-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 10:08 -------
There are 152 continuation objects, retained size 32 MB.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 15:53 -------
Now the heap usage after the final GC is 245 MB. Maybe this is because I didn't
clean the content before re-running the test.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 13:24 -------
(In reply to comment #6)
> The session is attached to SourceWrapper.lockedUris:

This is fixed in revision 570920, but I guess there's more.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From rfrovarp@apache.org  2007-11-30 15:00 -------
Here's what jmap -histo gives me after a day of running Lenya under some amount
of use. 130MB in int[] seems kind of big. Not sure how to force a gc. Need to
track down better methods of looking at usage on a production system.
Suggestions welcome.

Object Histogram:

Size    Count   Class description
-------------------------------------------------------
130975912       31216   int[]
61591232        552353  char[]
46645952        16659   byte[]
19199392        81286   * ConstMethodKlass
13899816        579159  java.lang.String
11170056        118103  java.lang.Object[]
9067280 93071   java.util.HashMap$Entry[]
7277952 303248  java.util.HashMap$Entry
6134720 188     org.apache.xpath.objects.XObject[]
5857880 81286   * MethodKlass
5047000 125054  * SymbolKlass
4679792 8576    * ConstantPoolKlass
3652968 8576    * InstanceKlassKlass
3646400 91160   java.util.HashMap
2850960 7585    * ConstantPoolCacheKlass
2384904 99371   java.util.ArrayList
1868704 6773    * MethodDataKlass
1498040 12528   java.lang.String[]
1032600 25815   org.apache.lenya.cms.site.tree2.TreeNodeImpl
1029760 32180   java.util.TreeMap$Entry
941088  29409   java.util.LinkedHashMap$Entry
940800  16800   org.apache.lenya.cms.publication.DocumentImpl
843784  11705   short[]
809072  9194    java.lang.Class
778824  32451   java.util.Date
752136  31339   java.util.Hashtable$Entry
680496  3565    int[][]
630480  7881    java.lang.reflect.Method
613536  25564   org.apache.lenya.cms.site.tree2.SiteTreeLink
580352  36272   org.apache.cocoon.xml.SaxBuffer$Characters
568272  13175   java.lang.Object[]
502256  31391   org.apache.cocoon.xml.ParamSaxBuffer
460656  28791   java.util.HashMap$KeySet
454720  11368   org.apache.lenya.cms.rc.CheckInEntry


-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240


rfrovarp@apache.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |




------- Additional Comments From rfrovarp@apache.org  2007-11-28 20:51 -------
I'm running at least r580365 from 2007-09-28, so after this bug was closed.
Lenya has been up for 13 days with max heap of 3096M. I am now getting:

java.lang.OutOfMemoryError: Requested array size exceeds VM limit

I know my ehcach.xml file is not set to eternal true. I'm not sure if I'm
missing other caches or if we have a leak. After a restart and initializing all
of my publications, my memory usage is about 515M.


It is something that needs to be tested again before release. 

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 09:47 -------
After a forced garbage collection it's at 200 MB.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 10:12 -------
The session is attached to SourceWrapper.lockedUris:

org.apache.lenya.cms.repository.SessionImpl
session of  org.apache.lenya.cms.repository.SourceNode
node of  org.apache.lenya.cms.repository.ContentSourceWrapper
value of  java.util.WeakHashMap$Entry
[37] of  java.util.WeakHashMap$Entry[1025]
table of  java.util.WeakHashMap
lockedUris of  org.apache.lenya.cms.repository.SourceWrapper
[2584] of  java.lang.Object[5121]
elementData of  java.util.Vector
classes of  org.apache.catalina.loader.WebappClassLoader
contextClassLoader of  java.lang.Thread [Stack Local, Thread]

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240


andreas@apache.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




------- Additional Comments From andreas@apache.org  2007-08-30 01:42 -------
It looks like the remaining objects are hold by continuations. That's a
different issue.

-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org


DO NOT REPLY [Bug 43240] - Memory leak

Posted by bu...@apache.org.
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG�
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=43240>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND�
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=43240





------- Additional Comments From andreas@apache.org  2007-08-29 10:03 -------
Phew ... the snapshot (after completion) indicates *lots* of Lenya objects.
There's definitely something wrong.

SessiomImpl:     402 objects (72 MB retained size)
SiteTreeImpl:    325 objects (75 MB retained size)
SiteTreeLink: 84'993 objects
DocumentImpl: 76'578 objects


-- 
Configure bugmail: http://issues.apache.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lenya.apache.org
For additional commands, e-mail: dev-help@lenya.apache.org