Please tell us about your environment:
- What OS?
- Using mod_jk with Apache (or other) in front? (If so, did you try
accessing Tomcat directly?)
- Always using a database for something? (If so, did you try with a
static resource?)
- Any firewalls in between?
When it fails, do in every involved
public interface MemoryLeak {
public void leak();
}
-
import java.util.Vector;
public class C implements MemoryLeak {
Vector v = new Vector();
public void leak() {
v.add(new Object());
}
}
-
public class MyLeakerServlet extends
You missed a /
url-pattern/servlet/TestServlet/url-pattern
Antonio Fiol
S.Latha Kamatchi wrote:
Yes Mr.Antanio,
As u said i used the invoker servlet...
My application files (Applet and bean files ) invokes servlets as
http://machineName:8080/webapps/myFolder/servlet/ServletName
I
Short answer is no.
Long answer is: Tomcat is free software. You can make it do what you
want. Simply patch it to your needs.
Antonio Fiol
Sonny Sukumar wrote:
Hmm ok. Is there any way to encrypt the passwords
stored in server.xml on top of that? I'll take all
the security I can get. :-)
Hi,
Does your app map all servlets? Or did you (lazily) use the invoker servlet?
The invoker servlet comes commented out in Tomcat 5. You have two options:
- Correct your app. (My advice, it makes your app more portable)
- Uncomment the invoker (Quick kludge, but it also works)
Yours,
Antonio
Hi,
I would establish two thresholds (possibly in %). Or maybe one threshold
could be spare and the other spare/4 or something.
max == busy seems too high for a simple warning. Apache web server
prints an error (not a warning IIRC) on the logs if MaxClients is reached.
Or maybe I dod not
Hi,
Threads may die... depending on your config.
You can use the Servlet#destroy() method. (Not Tomcat-specific)
Yours,
Antonio Fiol
Kenneth H. Cox wrote:
Problem: I call native code (Tcl) from my servlet. Tcl's thread model
forces me to run Tcl only on the thread that created the Tcl
Stephen Carville wrote:
Restarting tomcat clers this up.
That's good! :)
And bad. Every time I restart, Tomcat loses the state information for
established login sessions. Customer don't like that.
That (with a high probability) is because some objects they store in
sessions are
Filip Hanik (lists) wrote:
In any case could a cluster node that ran out of memory destroy the
entire cluster?
it shouldn't, it can temporary slow it down if the node that is down is
accepting connections and broad casting its membership.
I'm running a load test right now with the latest
wuhui wrote:
Antonio_Fiol_Bonn?,
can you give us something test data such as throughput, Average response time
etc. as a reference ?
No. At least not something reliable.
Typical average load time for our apps is between 500ms and 1000ms,
depending on the pages. Light pages go as low
chunked encoding (and ofcourse no content length)
in the second case?
Thanx in Advance,
Lukas sterreicher
Am 13.01.2004 21:13:51, schrieb Antonio Fiol Bonnn [EMAIL PROTECTED]:
My questions:
- how to disable chunked encoding even when there is no content length?
- why does tomcat not use
My questions:
- how to disable chunked encoding even when there is no content length?
- why does tomcat not use chunked encoding when sending mp3 files (and no content length is set)
and why does'nt it for ogg?
See RFC 2616. Using either the Content-Length header OR chunked encoding
is a
To: Tomcat Users List
Subject: Re: response.reset() and forward() ... problematic? DBCP related?
Hi!
Antonio Fiol Bonnn wrote:
My guess:
req and res are attributes of the Servlet, like in:
public class TestServlet extends HttpServlet {
private HttpServletRequest req;
private
Hello,
I know nothing about Content-Disposition. In fact, the only thing I know
is that we never needed/used it ;-)
Maybe it is too obvious, but I suggest you to try
response.setContentType() instead.
Anyway, I will try to get a code snippet from one of our apps a bit
later for you.
Hope that
Hi!
Sorry about my previous response: I missed the important point.
Here is what the RFC says about Content-Disposition (two fragments of
RFC 2616).
Anyway, the important part is that you SHOULD NOT send your file AND a
web page after it.
In your code snippet, you loop over the file, sending
on Tomcat 5.0.12
you can not do it with session replication, since null doesn't implement
java.io.Serializable,
request.getSession().setAttribute(name,null);
I can modify my code to make it work with code like that
- Original Message -
From: Antonio Fiol Bonnn [EMAIL PROTECTED]
To: Tomcat Users
to be serializable, if you want to cluster a webapp
Filip
- Original Message -
From: Antonio Fiol Bonnn [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Friday, November 14, 2003 10:11 AM
Subject: Session Replication problems on Tomcat 5.0.12
Hello,
I am having trouble with Tomcat 5.0.12
[A8FD8BD50E14477CC57E3EBA6F9591B8]
name=test; value=test
LastAccess=1068837178625
- Original Message -
From: Antonio Fiol Bonnn [EMAIL PROTECTED]
To: Tomcat Users List [EMAIL PROTECTED]
Sent: Friday, November 14, 2003 10:51 AM
Subject: Re: Session Replication problems on Tomcat
Message -
From: Antonio Fiol Bonnn [EMAIL PROTECTED]
To: Tomcat Users List [EMAIL PROTECTED]
Sent: Friday, November 14, 2003 11:53 AM
Subject: Re: Session Replication problems on Tomcat 5.0.12
Unfortunately, I cannot reproduce the error in a deterministic way. What
I know is that it sometimes
john d. haro wrote:
What is the load on your web servers?
Very low.
Could you repurpose a web server and
load balance the app server instead?
Web servers are light machines (uniprocessor, low memory, ...), while
app server is a heavyweight (two nice fast processors, 4Gb RAM, ...).
Web
+1 !! I liked the idea. I note it for future needs.
Antonio Fiol
Roman Fail wrote:
There's also an iptables workaround for port 80 on Linux that I am using in production and found very useful:
http://www-106.ibm.com/developerworks/linux/library/l-secjav.html
Roman
-Original
21 matches
Mail list logo
