tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Pae Choi" <paec...@earthlink.net>
Subject Re: Serious memory leak w/JSP pages?
Date Wed, 23 Jan 2002 18:17:12 GMT
If I remember correctly, this problem was caused either by the
limited heap size and/or the earilier J2SE SDK which supposed to
be fixed in the J2SE SDK v1.3.1_02 and later.

And Criag McClanahan was also addrssing the similar solution
before related to this kinda issue. Again, if I am not remembering
correctly, please accept my appologies. Otherwise, please confirm
and clarfy this issue one more time.


Pae



----- Original Message ----- 
From: "Michael Molloy" <mmncycles@yahoo.com>
To: "Tomcat Users List" <tomcat-user@jakarta.apache.org>
Sent: Wednesday, January 23, 2002 9:35 AM
Subject: Re: Serious memory leak w/JSP pages?


> I'll check it out.
> 
> Thanks
> --Michael
> 
> --- Simon Oldeboershuis <sol@outermedia.de> wrote:
> > Hi Michael,
> > 
> > you can change the session timeout in 
> >  tomcat/webapps/yourapp/WEB-INF/web.xml
> > 
> > We have been using OpenSTA it is for Windows only
> > but you can generate
> > the test by simply clicking in a browser. The tool
> > captures works as a
> > kind of proxy and records all request. You can
> > easily simulate how
> > different user would act on your website and run
> > multiple test parallel.
> > 
> > If you have an complicated HTML-Page with lots of
> > nested TABLE tags,
> > Netscape needs sometimes several seconds to render
> > the page... If the
> > page has more than 8 KByte normal modem users need
> > more than a second to
> > download it. On the other hand, you will need some
> > computing power to
> > create the output, depending on your application.
> > Lots of different
> > scenarios...
> > 
> > I read several times, that people use tomcat for
> > development. If it is
> > to slow you can still change to resin for realistic
> > testing and
> > deployment.
> > 
> > so long!
> >  Simon
> > 
> > PS: I would not expect 
> > 
> > 
> > Michael Molloy schrieb:
> > > 
> > > Hi Simon. You raise some good points. Our
> > requirements
> > > are that the system handle 150 concurrent users,
> > but I
> > > don't know if that means 150 requests every
> > second,
> > > every 2 seconds, every .10 a second, etc. Are
> > there
> > > any stress testing standards out there that anyone
> > is
> > > aware of?
> > > 
> > > Obviously I'm not trying to simulate all 150 on my
> > > Celeron 450, but I don't know how many is
> > realistic
> > > and how many is too many for my hardware.
> > > 
> > > As for the session length, what it's looking like
> > is
> > > that Resin (somehow) handles the default session
> > > settings (maybe it comes out of the box with a
> > shorter
> > > session timeout?) and Tomcat doesn't handle the
> > > default session settings nearly as well.
> > > 
> > > Thanks
> > > --Michael
> > > --- Simon Oldeboershuis <sol@outermedia.de> wrote:
> > > > Hi michael,
> > > >
> > > > I guess that JMeter is not reusing the same
> > session
> > > > and the session
> > > > timeout is set to 30 minutes. Have a look at
> > your
> > > > web.xml for the
> > > > session timeout value, e.g. tomcat 3.2.3
> > > >
> > > > <session-config>
> > > >  <session-timeout>30</session-timeout>
> > > > </session-config>
> > > >
> > > > if you have 50 threads, doing each 1 request per
> > > > second you have about
> > > >  50*60*30=90.000 sessions
> > > > after 30 Minutes. This could be a already a
> > memory
> > > > problem.
> > > >
> > > > Try to add something like
> > > >  request.getSession(true);
> > > > in your servlet to check if this really is your
> > > > problem.
> > > >
> > > > However, you should check for how many
> > concurrent
> > > > users your site is
> > > > aiming.
> > > >
> > > > Try to shorten the session timeout
> > > > so long!
> > > >  Simon
> > > >
> > > >
> > > > Michael Molloy schrieb:
> > > > >
> > > > > Testing Tomcat 3.3a on Mandrake 8.1 with Sun
> > JDK
> > > > > 1.3.1_02. 450 MHz Celeron w/192 MB RAM. Tried
> > the
> > > > same
> > > > > test with 4.0.1 with the same results.
> > > > >
> > > > > There appears to be a memory leak when calling
> > > > JSPs.
> > > > > For example, I start Tomcat standalone and use
> > > > JMeter
> > > > > on a different machine to create 50 threads
> > > > calling a
> > > > > JSP that simply prints out some environment
> > > > > information. After about 45 minutes, Tomcat
> > runs
> > > > out
> > > > > of memory & stops accepting connections.
> > > > >
> > > > > A servlet that uses a PrintWriter to print the
> > > > same
> > > > > information out to the calling browser can run
> > > > > indefinitely with no appreciable increase in
> > > > memory
> > > > > consumption.
> > > > >
> > > > > I have had the same result with other servlets
> > &
> > > > JSPs
> > > > > as well, and on Windows 2000 Server, also.
> > > > >
> > > > > Does anyone have any suggestions for fixing
> > this?
> > > > > Setting a page directive with session="false"
> > > > seems to
> > > > > prevent the memory leak, but we need to use
> > > > sessions
> > > > > in our application.
> > > > >
> > > > > I wonder if I'm missing something. I tried the
> > > > > identical JSP under Resin, and it ran for
> > hours
> > > > > without any increase in memory. If Tomcat
> > can't
> > > > serve
> > > > > JSPs for an hour under a moderate load without
> > > > running
> > > > > out of memory, that seems to be a pretty big
> > > > issue.
> > > > > Again, I hope I've missed something important
> > that
> > > > > will fix this.
> > > > >
> > > > > --Michael
> > > > >
> > > > >
> > __________________________________________________
> > > > > Do You Yahoo!?
> > > > > Send FREE video emails in Yahoo! Mail!
> > > > > http://promo.yahoo.com/videomail/
> > > > >
> > > > > --
> > > > > To unsubscribe:
> > > >
> > <mailto:tomcat-user-unsubscribe@jakarta.apache.org>
> > > > > For additional commands:
> > > > <mailto:tomcat-user-help@jakarta.apache.org>
> > > > > Troubles with the list:
> > > > <mailto:tomcat-user-owner@jakarta.apache.org>
> > > >
> > > > --
> > > > To unsubscribe:
> > > >
> > <mailto:tomcat-user-unsubscribe@jakarta.apache.org>
> > > > For additional commands:
> > > > <mailto:tomcat-user-help@jakarta.apache.org>
> > > > Troubles with the list:
> > > > <mailto:tomcat-user-owner@jakarta.apache.org>
> > > >
> > > 
> > > __________________________________________________
> > > Do You Yahoo!?
> > > Send FREE video emails in Yahoo! Mail!
> > > http://promo.yahoo.com/videomail/
> > > 
> > > --
> > > To unsubscribe:  
> > <mailto:tomcat-user-unsubscribe@jakarta.apache.org>
> > > For additional commands:
> > <mailto:tomcat-user-help@jakarta.apache.org>
> > > Troubles with the list:
> > <mailto:tomcat-user-owner@jakarta.apache.org>
> > 
> > --
> > To unsubscribe:  
> > <mailto:tomcat-user-unsubscribe@jakarta.apache.org>
> > 
> === message truncated ===
> 
> __________________________________________________
> Do You Yahoo!?
> Send FREE video emails in Yahoo! Mail!
> http://promo.yahoo.com/videomail/
> 
> --
> To unsubscribe:   <mailto:tomcat-user-unsubscribe@jakarta.apache.org>
> For additional commands: <mailto:tomcat-user-help@jakarta.apache.org>
> Troubles with the list: <mailto:tomcat-user-owner@jakarta.apache.org>
> 


--
To unsubscribe:   <mailto:tomcat-user-unsubscribe@jakarta.apache.org>
For additional commands: <mailto:tomcat-user-help@jakarta.apache.org>
Troubles with the list: <mailto:tomcat-user-owner@jakarta.apache.org>


Mime
View raw message