I'm not sure which class. In my experience, this is a development-only issue
that happens about once an hour for me. If you're experience it more, you're
probably restarting too much and should write more tests or use JRebel. ;-)

I have experienced it in production and wrote about my story on my blog:

http://raibledesigns.com/rd/entry/how_do_you_determine_a

Matt

On Wed, Sep 23, 2009 at 11:36 AM, ramzi khlil <ramzi.atv...@gmail.com>wrote:

> Matt,
>
> Can you give us more insights ? Which class in hibernate that causes this
> problem ?
>
> Ramzi
>
> On Wed, Sep 23, 2009 at 1:22 PM, Matt Raible <m...@raibledesigns.com>wrote:
>
>> It's not so much a memory leak as it is a problem that happens when you
>> restart your application several times. If you google around for this issue,
>> I believe it's caused by Hibernate internally, but I'm not certain.
>>
>> Matt
>>
>>
>> On Wed, Sep 23, 2009 at 10:50 AM, ramzi khlil <ramzi.atv...@gmail.com>wrote:
>>
>>> Hi,
>>>
>>> I'm investigating the memory leak that's causing OutOfMemoryException
>>> PermGen.
>>> I'm using jhat.
>>>
>>> Ramzi
>>>
>>>
>>> On Wed, Sep 23, 2009 at 10:34 AM, Matt Raible <m...@raibledesigns.com>wrote:
>>>
>>>> In the past, I've used WAPT (http://www.loadtestingtool.com/) and
>>>> JProfiler to detect memory leaks. WAPT works well in that in can simulate a
>>>> lot of concurrent users. JMeter, YourKit and BrowserMob are other good 
>>>> tools
>>>> to facilitate this.
>>>>
>>>> Matt
>>>>
>>>> 2009/9/23 Magnús Skúlason <magg...@gmail.com>
>>>>
>>>> Hi,
>>>>>
>>>>> I would like to know if anyone has used a profiler with an appfuse app,
>>>>> that is to investigate possible memory and resource leaks. If so it would 
>>>>> be
>>>>> great to get some pointers on how that is best/simplest done.
>>>>>
>>>>> Would I be best of by simply deploying the .war on a tomcat along with
>>>>> some profiler, I have seen a few that can simply be deployed to the
>>>>> container as well or does anyone know a better approach? What is the best
>>>>> profiler out there in terms of accurateness and simplicity.
>>>>>
>>>>> best regards,
>>>>> Magnus
>>>>>
>>>>>
>>>>
>>>
>>
>

Reply via email to