tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Tim Funk <funk...@joedog.org>
Subject Re: Regression? Tomcat 4.1.27 spiraling out of control!
Date Fri, 26 Sep 2003 23:32:23 GMT
IIRC when 4.1.28 comes out, (see the FAQ for a release date), I'm pretty sure 
it has some connector fixes which may help some people in your condition.

(Of course, I might be thinking of something different too)

-Tim

Jeff Larsen wrote:

> I just upgraded my production server to tomcat 4.1.27 
> running on Solaris 7 with java 1.4.2_01. I am using Ajp13
> connections to Apache 1.3.23 with mod_jk 1.2.0. I was
> running 4.0.4 on java 1.3.0 without incident.
> 
> It ran fine for a day and a half, but then over a period of
> about 10 minutes I got a flood of log messages for "Broken
> pipe", "Starting background thread" and finally when it exceeded
> my configured Ajp13Processor limit, "No processor available, 
> rejecting this connection". Also, tomcat pegged my CPU usage
> during this period. I bounced tomcat and now all is well with
> less than 10% cpu for tomcat.
> 
> The httpd logs show no evidence of a DOS attack. Besides, the
> tomcat application requires form based login.
> 
> My 4.0.4 setup never used more than 55 AJP processors. It was 
> (as is 4.1.27) configured for a max of 75 which it never reached.
> 
> I had the very same problem when I first deployed 4.0.3, but it
> got fixed in 4.0.4. Can anybody confirm this regression?
> 
> Google hasn't been very helpful on this one. Should I just
> revert to 4.0.4 and wait for a fix?
>  


Mime
View raw message