I've seen this problem in 3.3.m4...
I just upgraded to 3.3b1 on one box and haven't seen it yet, but haven't tested it as
thoroughly.
I'm really interested in a work around?
In addition to this problem I occasionally get a jsp page that isn't processed
returned.
I've seen this on two seperate m4 installs running different sets of code.
--aaron
On Thu, Jul 26, 2001 at 01:43:27PM -0500, Jason Koeninger wrote:
> I've seen this exact problem in 3.3m3, but I don't have any
> recommendations on working through it right now. ajp12
> has been better in this respect than ajp13, but the problem
> isn't gone. I'm planning to spend some time in August or
> September working through the mod_jk code to check on a
> number of issues I've had with it, this one being the most
> important. Until then, I'll be interested to see what kind of
> response your message gets.
>
> Best Regards,
>
> Jason Koeninger
> J&J Computer Consulting
> http://www.jjcc.com
>
> On Thu, 26 Jul 2001 19:31:51 +0100, Andrew Ormsby wrote:
>
> >Hi,
> >
> >I'm using Apache 1.3.20 and Tomcat 3.3m4 and I have the following
> >problem:
> >
> >I have a simple servlet which responds to GET requests by returning a
> >web page with some binary content. The precise content is selected by a
> >parameter, e.g: http://host/servlet/get?p=1
> >
> >Accessing the servlet from Tomcat standalone, everything works fine.
> >However, when using mod_jk (AJPv13) to access Tomcat from Apache, things
> >work well until after the server has been under heavy load.
> >
> >Once the server has been loaded up, I start getting bizarre behaviour:
> >Apache will return the output of the *previous* servlet invocation.
> >Given the sequence:
> > http://host/servlet/get?p=1
> > http://host/servlet/get?p=2
> >the first request will return some old result and p=2 will return the
> >result expected from p=1.
> >
> >From my servlet logs, I know that the p=1 request reaches the servlet
> >and the right content is generated (and I can confirm this by accessing
> >the servlet directly on port 8080, by-passing Apache). But Apache
> >resolutely returns the "old" content.
> >
> >This feels like some kind of buffering problem - perhaps in mod_jk, or
> >maybe in Apache itself. Has anyone else seen anything like this? Any
> >suggestions on how I can diagnose the problem?
> >
> >--
> >Andy Ormsby andy.ormsby .at. lexicle.com
> >
>
>
