Re: [Wrapper-user] Very Noisy Logs - "Previous control code"

SourceForge Headquarters 1320 Columbia Street Suite 310 San Diego, CA 92101 +1 (858) 422-6466

Mark,
Sorry for the delay.

The Wrapper's wrapper.dll file has as one of its jobs, the
interception of system signals.   It is not possible to make a call
back into the JVM from within the signal handler itself because a
second signal being thrown would result in the JVM crashing.

To work around this, the Wrapper stores the signal into a rotating
Queue and exits the signal handler.   A Java thread then polls this
Queue several times per second to see if any signals have been
trapped.    This Queue has a size of 10 currently.   It appears that
in your case, you are receiving 10 or more signals between the calls
from the JVM to poll for signals.

This code has existed as is for about 6 years, and this is the first
such problem I have heard of.   An easy fix is to increase the queue
size to something that would not cause problems, but I am wondering
why you are receiving so many signals in the first place.

According to the Windows docs, control code #4 is actually reserved.
Do you have any idea as to what might be sending this signal on your
system?  Is it reproduceable?    If so, could you enable
wrapper.debug=true and then send me the top of the wrapper.conf up
until the first few of these warnings?   I would like to see if that
offers any clues.

These are the declarations from the Windows Wincom.h header file.

#define CTRL_C_EVENT        0
#define CTRL_BREAK_EVENT    1
#define CTRL_CLOSE_EVENT    2
// 3 is reserved!
// 4 is reserved!
#define CTRL_LOGOFF_EVENT   5
#define CTRL_SHUTDOWN_EVENT 6

Cheers,
Leif

On Fri, Sep 10, 2010 at 11:49 PM, Wardell, Mark
<Mar...@co...> wrote:
> Hi
>
> I am using community edition 3.5.3 on windows (32 bit) and I gets "loads" of
> these in my logs. I have searched for this but haven't found anyone else
> reporting the same issues.
>
> WARN   | wrapper  | 2010/09/10 15:35:19 | Previous control code (4) was
> still in queue, overwriting with (4).
> Anyone else seen this and dealt with it?
>
> Thanks
> Mark