Re: [Wrapper-user] Classload problems with inner classes. Was Re: Classes unloaded too early on process termination

[Leif M. <le...@ta...>]

Chuck,
    What integration method are you using?  There were some minor 
changes to the
WrapperStartStopApp concerning how it decides when to shutdown the JVM.  
I don't
think that should have done anything to cause the problems you are seeing.
    There are not any other changes that I can think of that could 
possibly be causing
the problems you are seeing.
    Were there any other changes to your application other than the 
upgrade from 3.1.2
to 3.2.3?

    Could you post your full wrapper.conf and a wrapper.log file with 
wrapper.debug=true?
I may be able to get some clues from them.

Cheers,
Leif

Chuck Williams wrote:
> The problem appears to be worse than described below.  It appears that
> certain inner classes are not loaded properly.  I've only seen this
> since upgrading to wrapper 3.2.3 from 3.1.2 Does wrapper do something to
> affect class loading?  Has anybody else experienced problems accessing
> inner classes?
>
> Thanks for any help,
>
> Chuck
>
>
> Chuck Williams wrote on 10/29/2006 11:19 AM:
>   
>> Hi All,
>>
>> After recently upgrading from wrapper 3.1.2 to 3.2.3 an anomaly has
>> developed in my application.  When a normal kill signal is received on
>> linux (e.g., control-c when running from console) and
>> WrapperListener.stop() is called, occasionally classes are unloaded too
>> early, causing subsequent references to those classes in my shutdown
>> code to get NoClassDefFoundError.  One time an inner class in my stop()
>> method itself got this error!
>>
>> Can anyone explain or suggest a solution?  Has anyone else seen this?
>>
>> Thanks for any help,
>>
>> Chuck
>>