You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tomcat.apache.org by peter lin <pe...@labs.gte.com> on 2002/06/25 15:35:37 UTC
more on memory leak 4.1.5
hey Remy,
I did some more test and here is a bit more info. I tested jasper2 with
4.0.3 and the is NO memory leak there. In fact the memory profile is
more stable than jasper1. Also, the CPU usage profile is much more
stable with jasper2 with tomcat 4.0.3, so I doubt it is purely jasper2,
though it could be a combination of changes to catalina and jasper2.
if I have time later today or tomorrow I will try to use JProbe and
profile the memory usage. I'll post any useful data I find.
peter
> --- Remy Maucherat <re...@apache.org> wrote:
> > peter lin wrote:
> > > i just did some quick tests with scriptlet and jsp
> > tag pages. Both
> > > result in "out of memory" error after a couple
> > hundred hits using JMeter
> > > for benchmarking. Has anyone else seen this?
> >
> > I saw a few reports like that in tc-user, and I'm
> > investigating.
> >
> > So I'd like more details. Esp on the concurrency
> > level you're using. I
> > just noticed trouble if I was using high concurrency
> > levels with ab
> > (like 100), and I found that to be because each
> > HTTP/1.1 processor is
> > allocating buffers which are too big (each processor
> > allocates about 1M
> > worth of memory). With a lower concurrency level
> > (about 40-50), this
> > won't cause memory errors, even with the default VM
> > setting.
> >
> > Raising the VM memory limit provides a complete
> > workaround for that bug,
> > as the processors are only allocating a lot of
> > memory once, they are not
> > leaking it after that. I'm committing a change to
> > the constant used
> > (128K was really too much).
> >
> > I don't think that was your problem, though.
> >
> > Remy
> >
> >
> > --
> > To unsubscribe, e-mail:
> > <ma...@jakarta.apache.org>
> > For additional commands, e-mail:
> > <ma...@jakarta.apache.org>
> >
>
> __________________________________________________
> Do You Yahoo!?
> Yahoo! - Official partner of 2002 FIFA World Cup
> http://fifaworldcup.yahoo.com
>
> --
> To unsubscribe, e-mail: <ma...@jakarta.apache.org>
> For additional commands, e-mail: <ma...@jakarta.apache.org>
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
jasper2
Posted by peter lin <pe...@labs.gte.com>.
hey remy,
did this CVS commit
http://www.mail-archive.com/tomcat-dev@jakarta.apache.org/msg29253.html,
fix the bug where "FileNotFoundException" is thrown because the working
directory was deleted for force a refresh?
peter
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
Re: more on memory leak 4.1.5
Posted by Remy Maucherat <re...@apache.org>.
peter lin wrote:
>
> hey Remy,
>
>
> I did some more test and here is a bit more info. I tested jasper2 with
> 4.0.3 and the is NO memory leak there. In fact the memory profile is
> more stable than jasper1. Also, the CPU usage profile is much more
> stable with jasper2 with tomcat 4.0.3, so I doubt it is purely jasper2,
> though it could be a combination of changes to catalina and jasper2.
It's normal, since Jasper 2 recycles almost all of its objects.
> if I have time later today or tomorrow I will try to use JProbe and
> profile the memory usage. I'll post any useful data I find.
I've made a few tests with servlets, and it looks like the memory usage
is ok (well, at least, it doesn't leak memory; before my change
yesterday, each HTTP/1.1 processor was still allocating a lot of memory
when it was created).
Remy
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>