Do you have any native code components?
I had a serious problem with OutOfMemoryError: unable to create new
native thread errors too. They pretty much went away when I ditched
the Oracle OCI driver in favor of the thin driver.
BTW, the default linux kernel puts a hard limit of ~1k threads on
closer? Ummm, maybe, but still hideously broken. If you're using
BLOBs from Java, forget both Oracle and PostgreSQL - their JDBC drivers
are horrible. And if you want to lie awake at nights shivering,
desperately trying to clear your mind of a gastly, horrible vision -
take a look at the
We're seeing very uneven loadbalancing in our cluster of four tomcat
instances; it looks like #1 gets about 75% of the traffic. Does anyone
have any idea why this might be happening?
Our environment:
2 Apache 1.3.27 instances running mod_jk (all worker lbfactors set to
100)
4
Renaming the file in the CVS repository makes it impossible to build old
versions of your code. If you rename the file within CVS, then CVS
thinks that file has always been called the new name. Checking out old
tags will give you the file with the new name, not the old name. Most
likely this
From: Craig R. McClanahan [mailto:[EMAIL PROTECTED]]
If you're going to switch from https-http, you are totally wasting
your
time messing with https in the first place. It buys you nothing
except a
*perception* that you are more secure -- that is not the reality.
You keep repeating this
From: Craig R. McClanahan [mailto:[EMAIL PROTECTED]]
* The most common use case for wanting HTTPS-HTTP is to let you
log on with HTTPS so your password does not go across the
Internet unencrypted, but run the rest of the application on
HTTP. Having to reauthenticate again means they
Take a look at the Maverick MVC framework (http://mav.sourceforge.net). It is
Struts-like in concept, but you can run any model through an arbitrary pipeline of
XSLT (and several other) transformations in Cocoon-like fashion. The pipeline will be
efficiently connected with SAX events if
From: Justin Ruthenbeck [mailto:[EMAIL PROTECTED]]
That said, there's nothing inherently wrong with what he's doing and I
think it took quick to blame what he's seeing on this arrangement. To
make
a conclusive argument, I think we'd have to see the TrafficCop
object's
code.
To make a
I had to deal with this last week. Here's the important bit of
documentation:
http://java.sun.com/j2se/1.4.1/docs/api/java/net/InetAddress.html
Jeff Schnitzer
[EMAIL PROTECTED]
The Sims Online
-Original Message-
From: Aditya [mailto:[EMAIL PROTECTED]]
Sent: Saturday, December 07,
FYI, no it doesn't, it just causes the (Sun, at least) JVM to dump a
list of threads and their stacks to stderr. Note that it's the real
stderr, not System.err. This is a JVM feature. It can be done anytime
and is a *really* useful debugging feature.
Jeff
-Original Message-
From:
What are the non-obvious causes of OutOfMemoryError: cannot create new
native thread?
This is happening quite a bit on my Tomcat cluster (linux + sun
jdk1.4.1_01 + jboss/tomcat 4.0.4, no ejbs) under load, but the actual
thread count (by either ps or the thread dump) is relatively low, often
less
This is probably a more general java question:
My long-running Tomcat processes become huge. I have the max heap set
to -Xmx512m, yet after a day or so the virtual size of java reaches
upwards of 2GB, and the resident size sometimes exceeds 1GB. Shouldn't
I get OutOfMemoryErrors sooner? If
to a
contractual commitment, or is otherwise legally binding on elata
unless
confirmed by an authorised representative independently of this
e-mail.
Registered in England, number 1961405
-Original Message-
From: Schnitzer, Jeff [mailto:[EMAIL PROTECTED]]
Sent: 04 December 2002 12:59
Slightly offtopic:
Can anyone recommend a profiler to use with webapps under Tomcat? Is
everyone happy with OptimizeIt? What about JProfiler? JProbe?
I want to tune a single webapp running under JBoss/Tomcat (no EJBs).
Thanks in advance,
Jeff
--
To unsubscribe, e-mail: mailto:[EMAIL
I understand why the http connector would get Broken pipe messages -
users hitting stop on their browser. But why does the Ajp13Processor
produce Broken pipe messages?
I'm getting a lot of this in my logs:
java.net.SocketException: Broken pipe
at
From: Price, Erik [mailto:[EMAIL PROTECTED]]
If they are overkill for webapps, what sorts of situations would
benefit
from them?
(Not asking rhetorically, I'm honestly curious about when to use EJB
since
I know nothing about them.)
Applications which are very transaction-oriented will
to the request,
the other browsers don't.
Tomcat doesn't know about the special meaning of
the anchor (#123) as it as client side thing.
-Original Message-
From: Schnitzer, Jeff [mailto:JSchnitzer;maxis.com]
Sent: Tuesday, November 05, 2002 7:37 AM
To: [EMAIL PROTECTED
This is the most obscure bug I have ever seen:
Take a look at the very simple JSP at the end of this message. If you
include go=true as a query parameter, it issues a redirect to itself
like this:
redir.jsp?foo=bar#123
If you use the redirect from IE6, Tomcat displays foo bar#123 rather
than
Hi! I'm using mod_jk to mount the root level of a webapp:
JkMount /* balancer
Is there a way to exclude some directories? In particular, I'd like
Apache to serve all content in /static.
I've tried:
Alias /static /my/path/to/static
...but that doesn't seem to work :-(
Yes, I know I
[mailto:[EMAIL PROTECTED]]
Sent: Tuesday, October 15, 2002 5:46 PM
To: Tomcat Users List
Subject: Re: Excluding directories from mod_jk mount
On Tue, Oct 15, 2002 at 05:36:00PM -0700, Schnitzer, Jeff wrote:
Is there a way to exclude some directories? In particular, I'd like
Apache to serve
Hi! I'm bewildered by the array of connector options. I'm trying to
hook up a cluster of Tomcat 4.0.4 instances (running in JBoss) behind
Apache 1.3, all running on Linux. I'm looking for a stable,
production-quality option.
It would be really really helpful to me and probably everyone else
accordingly.
Ralph Einfeldt
Uptime Internet Solution Center GmbH
Hamburg, Germany
Hosting, Content Management, Java Consulting
http://www.uptime-isc.de
-Ursprüngliche Nachricht-
Von: Schnitzer, Jeff [mailto:[EMAIL PROTECTED]]
Gesendet: Mittwoch, 14. August 2002 01:58
I have a conceptual question about load balancing Tomcat4.0.4 instances.
I'm building a *very* high traffic site. Traditionally this company handles
high loads with layering like this:
Non-sticky load balancer
Several NES instances
Many WebLogic instances
This works because WebLogic session
23 matches
Mail list logo