tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rainer Jung <>
Subject Re: [SPAM] RE: Application stops responding, jk worker in error state
Date Sat, 19 Jun 2010 08:38:22 GMT
On 19.06.2010 03:46, Neil Aggarwal wrote:
> Rainer:
>> Hmmm, errno 11 is often EAGAIN. What platform are you using
>> (OS, version)?
> I am using CentOS 5.5 freshly installed, not an upgrade.
> I tried that as a fix to this problem.  I wiped the server
> clean and reinstalled everything from scratch.  It was on
> CentOS 5.4 before and I had the same behavior.
>>> What is really strange is the JK Status page says -48
>> connections.  It seems
>>> like that number should always be positive or zero but not negative.
>> That should be unrelated (and yes: it's not OK).
> It seems to me if jk is confused about the number of
> backend connections, it may not be connecting correctly
> to Tomcat.  That seems like it would cause the behavior
> I am experiencing.

No, the total number of connections is shown only for convenience 
(monitoring etc.). The real process pool is local for each Apache 
process. there seems to be something wrong when tracking the total 
number via shared memory but that shouldn't influence in any way the 
usual functionality. It has only been added very recently.

What *is* a problem is EAGAIN while reading from the socket. We don't 
expect that and handle it as an error, though strictly speaking it is 
only a temporary error condition. i'll see whether I can provide a patch.

would you mind opening an issua in Bugzilla, attaching your log snippet 
and possibly your (the property dump from the status 
worker is a little harder to digest, because it also shows all the 

You can also open a second issue concering the obvisouly wrong 
connection count shown in the status worker.

Thanks and sory for the trouble


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message