Re: [Wrapper-user] JAVA 1.6 support

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

Nitin,
There are no known problems with that combination of software.   If
you could send me a wrapper.log with wrapper.debug=true set that shows
the events that lead up to this happening it would enable me to give
you some useful suggestions.

Also, please describe the timing of the creation of these additional
Java processes.  Are they happening when you stop the Wrapper?  Or
when a JVM restart is triggered by the Wrapper?

I also urge you to try out the 3.3.5 release.  There have been many
issues fixed and improvements made.  The 3.2.1 release is already a
couple years old.

Cheers,
Leif

On Wed, Jun 24, 2009 at 5:28 PM, Nitin Jain (XIS)<Ni...@xc...> wrote:
>
> Thanks Leif for the clarification.
>
> Our application doesn't spawn child threads. Can you please confirm that this is not a compatibility issue between release 3.2.1/Java 1.6/Solaris? I will explore other areas then.
>
> thanks
>
> -----Original Message-----
> From: Leif Mortenson [mailto:lei...@ta...]
> Sent: 24 June 2009 13:45
> To: wra...@li...
> Subject: Re: [Wrapper-user] JAVA 1.6 support
>
> Nitin,
> The problem I was talking about was actually introduced in 3.3.0 and
> then fixed in the 3.3.1 release.  If the Wrapper decided that the JVM
> was frozen it was supposed to kill the JVM and once it was confirmed
> to be gone, then launch a new JVM.  The problem was that there was a
> bug in the state engine where the Wrapper was not waiting to confirm
> that the "killed" JVM was really gone.   In some cases, it would be
> left around.
>
> If you are using 3.2.1, you would not be encountering the above bug.
>
> The JVMs must be getting caused by something else.   Like I said in my
> last email, seeing a debug wrapper.log would be helpful (set
> wrapper.debug=true).
>
> Does your application spawn child processes at all?  I have seen user
> applications in the past which spawn child JVMs and then fail to clean
> them up.
>
> Cheers,
> Leif
>
> On Wed, Jun 24, 2009 at 3:13 PM, Nitin Jain (XIS)<Ni...@xc...> wrote:
>>
>> Hi Leif,
>>
>> We are seeing multiple Java processes. As you said the fix was in 3.3.0 but we are using 3.2.1, could this be an issue?
>>
>> Regards,
>>
>> Nitin
>>
>> -----Original Message-----
>> From: Leif Mortenson [mailto:lei...@ta...]
>> Sent: 23 June 2009 22:45
>> To: wra...@li...
>> Subject: Re: [Wrapper-user] JAVA 1.6 support
>>
>> Nitin,
>> Which processes are you seeing multiple copies of?  The Wrapper or Java process?
>>
>> What is the timing of the creation of the new processes?
>>
>> There is not anything I can think of that would cause Java 1.6 to
>> perform differently than Java 1.4.
>>
>> We did fix a problem for the 3.3.1 release.   Frozen Java processes
>> were not always being killed correctly resulting in Zombie Java
>> processes.   I am pretty sure that but had been introduced in 3.3.0
>> however.
>>
>> I will try to give you a better answer once I hear back from you.
>>
>> Cheers,
>> Leif
>>
>> On Wed, Jun 24, 2009 at 12:57 AM, Nitin Jain (XIS)<Ni...@xc...> wrote:
>>> Hi,
>>>
>>>
>>>
>>> We are using wrapper version 3.2.1. Now we have upgraded our application
>>> from Java 1.4 to Java 1.6 without upgrading the wrapper version.
>>>
>>>
>>>
>>> Sometimes it happens that multiple instances of a particular process start
>>> running concurrently. Could this be a compatibility issue as we are now
>>> using Java 1.6?
>>>
>>>
>>>
>>> Thanks
>>>
>>>
>>>
>>> Nitin
>>>
>>>
>>>
>>>
>>>
>>> __________________
>>>
>>> Xchanging
>>> 271, Udyog Vihar,
>>> Phase II, Gurgaon - 122 006
>>> Haryana, INDIA.
>>> Mobile:                +91 (0) 9810321221
>>> Switchboard:        +91 (0) 124 4339333, Ext: 230
>>>
>>> Email:                  ni...@xc...
>>> Website:              www.xchanging.com