Re: [Wrapper-user] EXCEPTION _ACCESS_VIOLATION

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

Mark,
What does the wrapper.log look like just before the crash?  I am
wondering if the Wrapper is thinking that the JVM is frozen and
attempting to restart it.  If the native code is not written correctly
it could be leading to a crash as the JVM tries to shutdown.
I ask because you mentioned that making the ping timeout longer
helped.

Could you enable wrapper.debug=true and try this again.  Then
post back or send the resulting log files (from a couple minutes before
the crash) to the list or to me directly if it has private data.

Cheers,
Leif

no spam wrote:
> Hi Leif,
>
> The wrapper process seems to have problems restarting the jvm 
> process.  In fact after this crash there will be several other crashes 
> noticeable in the event viewer and also in the wrapper log.
>
> Right the Dr Watson output is very confusing.  I've been running a 
> test since last Thursday without the wrapper and everything seems to 
> be running great.  I believe that my last test ran better when I upped 
> the ping time in the wrapper config.  This process runs in parallel 
> with a second JVM process that hovers around 90-95% cpu, could this be 
> an issue?
>
> Yes I'm using JIntegra to talk to another vendor's objects (IPV).  I 
> should know a bit more in a few days if things keep running.  This is 
> really some interesting software here at CNN.  This code controls all 
> the low resolution encoders for the newsroom.   It has to be 100% 
> stable which is why I have been load testing for the past few weeks.
>
> Thanks for your prompt reply.  I really want to be able to run this 
> via the wrapper as it's so convenient.
>
> Mark