I think you should post this to the [EMAIL PROTECTED]
mailing list. to get it to work with JMeter, you'll need to add auth
manager to your test plan.
add an entry in the auth manager
url - leave blank
username - yourUser
password - yourPassword
The documentation for it is here.
first we need to know what kind of hardware you're using. there are
articles on tomcat's resource page.
http://jakarta.apache.org/tomcat/resources.html
tuning the performance should take into consideration what kind of
load you expect, so if you don't have those written down as
requirements, I
has anyone ever tried to use url-pattern to filter for basic auth
beyond /* all? If I do this, it forces all requests to my servlet
to authenticate.
servlet
servlet-nameAuthServlet/servlet-name
servlet-classtest.AuthServlet/servlet-class
/servlet
servlet-mapping
of
these.
-Original Message-
From: Peter Lin [mailto:[EMAIL PROTECTED]
Sent: Monday, November 08, 2004 1:57 PM
To: tomcat-user
Subject: url-pattern with Basic Auth
url-pattern/AuthServlet?*param1=account1*/url-pattern
I tried it and cross-referenced the servlet spec. It doesn't
, Shapira, Yoav [EMAIL PROTECTED] wrote:
Hi,
It's SRV.11.2 in the Spec, and org.apache.tomcat.util.http.mapper.Mapper
in the Tomcat source code.
Yoav Shapira http://www.yoavshapira.com
-Original Message-
From: Peter Lin [mailto:[EMAIL PROTECTED]
Sent: Monday, November 08
if you're using hardware load balancer like cisco localdirector, I
would setup the load balancer to direct the traffic based on
sessionid.
this way, you don't need to use keep alive. when you say 8K
simultaneous users, what does that translate to in terms of concurrent
requests per second? An
connections? Does it really keep 1 thread open for
each keep-alive? this seems VERY unnecessary
Regards
Andrew
On 29.10.2004, at 17:02, Peter Lin wrote:
if you're using hardware load balancer like cisco localdirector, I
would setup the load balancer to direct the traffic based
whether
keepalive should be disabled or not.
peter
On Fri, 29 Oct 2004 17:44:09 +0200, Mladen Turk [EMAIL PROTECTED] wrote:
Peter Lin wrote:
you don't need to use keepalive. generally, in a load balanced setup,
keepalive is disabled because the load balancer is already making sure
the user
times) per request.
Andrew
On 29.10.2004, at 17:53, Peter Lin wrote:
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
a dedicated image server.
I hope that helps
peter
On Fri, 29 Oct 2004 19:24:07 +0200, Andrew Miehs [EMAIL PROTECTED] wrote:
On 29.10.2004, at 19:08, Peter Lin wrote:
if you're looking for better client performance I would explore other
areas first.
1. use gzip compression - this can reduce
it depends on which wap gateway it is.
the older wap gateways had a page limit, so the solution to that is to
break pages into smaller chunks. It's hard to tell if the problem you
see is related to that. back when i worked on WAP in 99, we figured
out that openwave's WAP gateway had a limit of
My feeling on this is, doing a sync on a data bean in this specific
case is not worth it. If the transaction is complex, which this case
isn't, use something like java transaction API.
doing either a subquery or the prepared statement example is a far
better way. doing a sync unnecessarily is a
distributed locking of session data implies you need something like a
distributed cache. If that is the case, maybe you should look at
Coherence.
distributed locking is meant to solve a specific problem, so unless
you really need to synchronize distributed objects like a distributed
cache, I
am I missing something, but looks like you're trying to build some
kind of web cache. why not use Hibernate or something that already
does caching for you instead?
the only time I can see a need to sync, is if the request contains
data that requires a transaction. Which in that case, you're
on a block of
code inside a method works.
Thanks,
Malia
-Original Message-
From: Peter Lin [mailto:[EMAIL PROTECTED]
Sent: Thursday, September 30, 2004 10:26 AM
To: Tomcat Users List; [EMAIL PROTECTED]
Subject: Re: method level synchronization doesn't work
am I missing
I would second Filip's comment. locking tomcat in this fashion is not
advised, unless you have solid experience with it. It's not as easy
as one would think.
peter
On Thu, 30 Sep 2004 10:36:34 -0500, Filip Hanik - Dev
[EMAIL PROTECTED] wrote:
sounds like your system needs a little rethinking,
Swag, makes me think Monster garage!
maybe Jesse james can team up with apache :)
peter
On Thu, 30 Sep 2004 12:07:27 -0400, Shapira, Yoav [EMAIL PROTECTED] wrote:
Hi,
For those of us who really like Tomcat (or have friends who really like
Tomcat): http://www.cafepress.com/meepzor/386851.
this magic query, it's not quite as simple as let the
databse handle it, is it?
-Original Message-
From: Peter Lin [mailto:[EMAIL PROTECTED]
Sent: Thursday, September 30, 2004 7:42 PM
To: Tomcat Users List; [EMAIL PROTECTED]
Subject: Re: method level synchronization doesn't work
that would all depend on what kind of apps you are building. If you
don't need EJB and only want to take advantage of JMS, then I would
just use openJMS or Joram with tomcat. hope that helps
peter
On Wed, 29 Sep 2004 16:28:31 -0400, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote:
Hi,
We are
I thought Oracle fixed that bug in their connectionPooling driver back
in 2001. Am I missing something? Are you using some other driver to
create a pool of jdbc connections to Oracle? sorry for the question if
it's already been answered in earlier messages.
I've used classes112.zip jdbc driver
-Original Message-
From: Shilpa Nalgonda [mailto:[EMAIL PROTECTED]
Sent: Tuesday, September 28, 2004 2:09 PM
To: Tomcat Users List; Peter Lin
Subject: RE: Tomcat 4.1 Connection Pooling...
I am using Oracle 8i, and Oracle drivers i use is Oracle111.jar, and
use
DBCp for connection
just in case, here's a link to Oracle's developer jdbc section
http://www.oracle.com/technology/sample_code/tech/java/sqlj_jdbc/files/jdbc30/index.html
just make sure the apps are calling java.sql.Connection.close() when
they are done. the PooledConnection underneath will pass the
connection
them to the pool?
-Original Message-
From: Shilpa Nalgonda [mailto:[EMAIL PROTECTED]
Sent: 28 September 2004 19:43
To: Tomcat Users List; Peter Lin
Subject: RE: Tomcat 4.1 Connection Pooling...
I have updated with classes12.jar, but still i have this problem. But
Peter
what ii
I don't think so, but it's got JMX support. what kind of operations
are you thinking of?
if you're talking about deploying a webapp, the management stuff
handles that. If you're talking about monitoring, you can use JMeter
2.0 to monitor tomcat 5.x
peter
On Tue, 28 Sep 2004 12:40:58 -0700
I would second that advice. the biggest factor for performance with
XML is having a ton of memory. If the application doesn't use XML, I
would suggest running some benchmarks with a variety of settings to
see which works best.
good luck
peter
On Thu, 23 Sep 2004 15:03:01 -0400, Fournier,
you could always read the performance article I wrote, which is listed
on the resources page.
http://jakarta.apache.org/tomcat/resources.html
I also have a follow up article posted on Jmeter's wiki.
http://wiki.apache.org/jakarta-jmeter/JMeterLinks
peter
On Thu, 23 Sep 2004 16:40:04 -0400,
are you using a single system as the client or multiple? I've been
able to simulate upward of 100 threads in jmeter against tomcat for a
static file using my Gateway laptop which is a 1.4ghz PentiumM with
1Gb RAM. beyond 100 threads, the throughput tends to go down. I don't
remember the exact
just out of curiousit, you're running Tomcat on windows server right?
in the even you're using windows professional, it is limited to 10
inbound connections, which is easy to overlook and forget.
I've easily gone over 1000req/sec for static files. For real pages
that hit the database, that's
,
but with better hardware I could manage that.
Any way, where could I look this inbound limit in a Windows system?
Maybe that is a hint to find de final answer.
Thanks,
Jorge Sopena
Peter Lin wrote:
just out of curiousit, you're running Tomcat on windows server right
so does this mean Intel should donate some hardware to Apache for
suggesting users buy more hardware :)
sorry, couldn't resist.
if you provide more information about what the page does (without
revealing sensitive info), you'll get more detailed response.
Otherwise, we're just shooting darts
have that, you should be able to predict the performance
fairly accurately.
having said that. it doesn't hurt to have more hardware :)
peter
On Tue, 14 Sep 2004 19:38:31 +0200, Mladen Turk [EMAIL PROTECTED] wrote:
Peter Lin wrote:
so does this mean Intel should donate some hardware
out of curiousity, how many times did you run the test?
doing transaform inside Tomcat will be slow the first couple of times.
what I've done in the past is run it for a couple hundred times, then
take the measurement for real.
peter
On Mon, 13 Sep 2004 07:02:06 -0700 (PDT), Michal Sg [EMAIL
wow, that just seems wrong. Have you tried other jdbc drivers? I know
from first hand experience with Oracle's jdbc driver, w/o pooling it's
minimum of 100ms to get connection. With pooling, it's usually less
than 5ms.
peter
On Fri, 10 Sep 2004 18:56:42 +0200, Henrik Rathje [EMAIL PROTECTED]
this really be a driver issue?
which connectionpool did you use while doing the measurements for your
'so you want high performance' paper?
regards, henrik
On Fri, 10 Sep 2004 12:02:13 -0500
Peter Lin [EMAIL PROTECTED] wrote:
wow, that just seems wrong. Have you tried other jdbc drivers
it's an easy mistake to make. the usual trick to benchmark and
performance testing is to prime the server a bit. we've all made
that mistake at some point.
peter
On Fri, 10 Sep 2004 19:35:43 +0200, Henrik Rathje [EMAIL PROTECTED] wrote:
Hi,
i did not change any settings, no restart and no
it is not on by default due to spec issues. for tomcat to be strictly
compliant, by default it should not strip the extra carriage returns.
If you search the mailing list back to 2001-2002, you see there was
lots of discussion about it. the funny thing is, it also makes it easy
to tell when a
well I don't consider that an security issue. just because you know
someone is using jsp tags, it doesn't mean you know how the whole
architecture works. The only thing it tells a competitor is that it is
feasible to use jsp tags.
beyond that, all the important and interesting stuff is what makes
you're option is to back port the status servlet to tomcat4. if you
manage to do that, it should work I would think :) If you do, I will
gladly use Jmeter's tomcat5 monitor to test it.
peter
On Tue, 24 Aug 2004 10:22:57 -0700 (PDT), Prasheel Vemulapalli
[EMAIL PROTECTED] wrote:
I want to
oh, I didn't realize the new 4.1.31 includes back port of the new
status servlet in tc5.
or am I mis-reading that?
peter
On Tue, 24 Aug 2004 20:16:08 -0700, Bill Barker [EMAIL PROTECTED] wrote:
Shapira, Yoav [EMAIL PROTECTED] wrote in message
news:[EMAIL PROTECTED]
Hi,
Does Tomcat
I would recommend downloading one of profiling tools like borland
OptimizeId or JProbe to profile the webapp. Use JMeter to simulate
load and watch what happens in the profiler. if you suspect memory
leaks, look at the heap and how it grows. if you search the archives,
you'll see plenty of
What kind of weight based load balancing are you thinking of? the
type of load balancing I've used is based on one of several.
1. ping time
2. preset load factor - usually based on hardware
3. system load - the weight is calculated using a hash of memory and cpu usage
in situations where all
sounds like you already have an idea of how to implement it and just
want feedback on what parameters it should consider.
from past experience, many people use a simple Hash when all the
system have the same hardware. When it is not, I would think memory
and network IO would be the biggest
you can always try JFluid, which is an experimental VM from sun that
has some cool profiling features.
peter
On Fri, 13 Aug 2004 09:05:51 -0400, Nandish Rudra [EMAIL PROTECTED] wrote:
JProbe is also a nice tool for tracking JVM behaviour. You may want to look
into it.
Nandish Rudra
ECI
I thought the only way to run a 64bit VM was on solaris, which means
you have to use the Sun JVM. Does IBM offer a 64bit JVM now?
peter
On Fri, 13 Aug 2004 11:55:44 -0500, Caldarale, Charles R
[EMAIL PROTECTED] wrote:
From: Dale, Matt [mailto:[EMAIL PROTECTED]
Subject: [OT] - JVM Max Heap
sweet! man that is a nice toy. I'd check with IBM to see if they have
a beta JVM you can try out. They must be working on one.
peter
On Fri, 13 Aug 2004 18:00:19 +0100, Dale, Matt [EMAIL PROTECTED] wrote:
Good point, its an IBM machine with 16 P4's and 12gig of ram. So its looking like
now if only I had a system to load test 32 vs 64bit JVM, I'd post some numbers.
peter
On Fri, 13 Aug 2004 12:11:13 -0500, Caldarale, Charles R
[EMAIL PROTECTED] wrote:
From: Peter Lin [mailto:[EMAIL PROTECTED]
Subject: Re: [OT] - JVM Max Heap Size on Linux
I thought the only way to run
I've used JProfiler in the past and I found it somewhat unreliable,
since it is pretty heavy weight. I should say it was based on a half
dozen tests using JProfiler and not a scientific evaluation.
it was the free eval version a couple years back. I find optimizeIt
more reliable for me and a
having used SunOne in the past, I would agree 2-5x faster is not
likely or even possible. SunOne is a combination of the old Netscape
code, jvm and the reference implementation of the servlet spec.
In practice, it is no different than servlet containers that use
native library for handling
this is most definitely an issue with glibc, since I've seen it in the
past on my machine. I've had it happen with Sun jvm and IBM jvm for
redhat in the past.
your best bet is to search the redhat mailing list for solutions to the problem.
peter
On Sun, 01 Aug 2004 10:54:51 -0700, David Rees
as others have stated, the only way to know is to stress test your
app. even without running a test, I can tell you IIS5 running on a
dual CPU box cannot handle 500 concurrent requests for dynamic pages.
500 concurrent requests for static files isn't a problem.
Depending on the kind of
running on the same machine.
-Original Message-
From: Peter Lin [mailto:[EMAIL PROTECTED]
Sent: Wednesday, July 28, 2004 17:53
To: Tomcat Users List
Subject: Re: JVM memory size changing dramatically
as others have stated, the only way to know is to stress test your
app. even
In case I gave you a rosy picture of websphere and MQSeries, there are
known bugs with MQSeries jms client. It is one of the most scalable
messaging systems, but it is not without bugs like all software. a
good friend of mine has encountered several (3) bugs with IBM's jms
client, which required a
as a general for myself, I always explicitly close the connections, to
make sure the connection gets returned to the pool.
peter
On Tue, 13 Jul 2004 12:37:53 -0500, Bliesner, Christopher P
[EMAIL PROTECTED] wrote:
Hi-I'm using Tomcat 4.1.12 and I have have it running great, thanks. I
have my
I remember back in tomcat4 several people discovered that long JSP
pages would fail because the file exceeded the limit in javac, or
using include directive resulted in a servlet with deeply nested
try/catch blocks.
what is the exact error you are getting?
peter
On Wed, 7 Jul 2004 10:09:47
have you tried handling it this way?
1. compile the jsp on a staging server
2. tar up the files
3. upload to the production servers
4. untar
5. touch jsp files
6. touch compiled classes in the working directory
I used that approach in the past to handle updates to the production servers.
peter
Over the last 2 years, I've done a lot of testing with tomcat4 and 5
on linux, windows and solaris.
A clean tomcat install on a X1 rackmount running solaris 9 does not
exhibit this kind of behavior. where there's no load on the system,
the java process is typically below 1%. As other's have
just in case. you did install the solaris patch for the JVM right?
if not, it will cause odd behavior
peter
On Fri, 25 Jun 2004 11:53:33 -0400, Guy [EMAIL PROTECTED] wrote:
Please! Any ideas?
Thanks,
Guy
-Original Message-
I have a Sun box with Solaris 9.
Tomcat
: Peter Lin [mailto:[EMAIL PROTECTED]
Sent: Friday, June 25, 2004 11:58 AM
To: Tomcat Users List
Subject: Re: Tomcat exits without any known reason
just in case. you did install the solaris patch for the JVM right?
if not, it will cause odd behavior
peter
On Fri, 25 Jun 2004 11:53:33
by default, the max threads is set to 150 in server.xml. Is your
webserver really getting that many active requests at one time? I've
worked on some large sites and most of the time, the peak concurrent
requests is usually less than 10. this is on some big Sun boxes.
just because there are 150
here's part of what Tomcat5 status servlet displays
Max threads: 150 Min spare threads: 25 Max spare threads: 75 Current
thread count: 25 Current thread busy: 3
Max processing time: 371 ms Processing time: 0 s Request count: 5
Error count: 1 Bytes received: 0.00 MB Bytes sent: 0.02 MB
As you can
if you're using tomcat in standalone mode, I would recommend
commenting out the jk stuff all together. I noticed your minProcessors
is 100 and max is 500. You must get some serious load, or the usage
pattern is such that an unique user session has lots of requets.
good luck.
peter
On Wed, 23
from my own experience, this is caused by something in the application
and not in Tomcat. the only way you're going to be able to figure it
out is to run Tomcat in OptimizeIt and watch it.
there's no easy way to track down the cause. It is achievable, just
requires lots of patience.
peter
On
There's a couple of different ways to handle this.
1. do not do it in real time. this is the easiest solution, but it
means a human has to be the one who figures who gets what class.
2. use JMS to update each user's session and make it so that once the
class has no more entries, no one else can
you don't necessarily need to lock or synchronize in a first come
first serve mode. simple put the order into a queue and generate a
temporary order id. Once the order goes through, the real order ID is
emailed to the user.
peter
On Mon, 21 Jun 2004 14:47:52 -0700, Mufaddal Khumri
[EMAIL
this might be a bit off topic, but if you want to profile your app in tomcat, Sun
Research has an experimental VM called JFluid. http://research.sun.com/projects/jfluid/
it is basically the jdk1.4.2_03 vm with hooks for profiling. it might help track down
GC and memory issues for those who
this has been mentioned countless times on the mailing list and I have tons of numbers
comparing client to server in my article on the resources page of tomcat.
if you want hard numbers, I would suggest look at the article, or run some stress
tests on your own apps. a quick test will give
ahh gotta love benchmarks. the only valid benchmark is your own application, which
you've tuned.
all other cases are seriously error proned or not applicable to real applications.
peter
Eric VERGNAUD [EMAIL PROTECTED] wrote:
le 16/06/04 21:50, Matt Bathje à [EMAIL PROTECTED] a écrit :
I
yup, you sure can if you're willing to subject the users to the annoying message. What
some people do is get a dedicated image server and setup https on it. that way, you
don't get the annoying warning and you don't impact the webserver doing the real work.
peter
Gabi [EMAIL PROTECTED]
I would like to add the following. Probably the biggest factor in webserver
performance is the concurrent requests. 3000 concurrent users who request 1 page every
5 minutes doesn't mean anything and most likely won't generate much load.
for the sake of illustrating the problem, say the
I wish I could help you here, but I haven't used it. I know some people on the mailing
list have. Hopefully one of them will respond. You probably should post a message to
MC4J mailing list for asssitance.
peter
Dale, Matt [EMAIL PROTECTED] wrote:Hi,
I downloaded this and it seems to have
the good news is there's nothing much you can do to optimize tomcat's SSL. The bad
news is SSL is CPU intensive. the most reliable way to improve SSL performance is to
get a hardware accelerator. People tend to disagree on this, but I am biased in favor
of using hardware acceleration.
If I am not mistaken, coldfusion now uses jsp tags. In fact, if you look at the
output, you'll notice tons of extra /r/n in the generated HTML code.
you're best bet is to look at the version of JRun you're using and which spec version
it supports. either tomcat4 or 5 should be compatible.
if you can upgrade to tomcat5, I would recommend it. the reason I suggest this is TC5
has the new status servlet, which will tell you how much heap is actually in use
currently.
the JVM will not release memory back to the OS that is true. in terms of performance
the biggest indicator of
what kinds of issues
you might see.
peter
Michael Duffy [EMAIL PROTECTED] wrote:
Thank you, Peter Lin. I wasn't aware of the new
status servlet. I'll look into an upgrade.
One question: any problems reported with installing
Tomcat 5.0.x as a service on Windoze servers? I
believe
You might want to look at the VM performance numbers in my Performance article. The
link is listed on the tomcat resources page. Generally, tweaking the generations
takes time.
for something, like XML, tweaking generation ratio doesn't help.
peter
Michael Duffy [EMAIL PROTECTED] wrote:
/articles/performance.pdf
I apologize for overlooking it for all this time.
I'll be sure to go through it right away. Thanks -
%
--- Peter Lin [EMAIL PROTECTED] wrote:
You might want to look at the VM performance
numbers
in my Performance article. The link is listed on
the
tomcat
For those who use JMeter to load test Tomcat, now
there's a distribution graph for JMeter. I just
checked it in and it should be in the nightly build.
peter
__
Do you Yahoo!?
Friends. Fun. Try the all-new Yahoo! Messenger.
I've seen this kind of behavior happen in the past because the HttpSessions are not
getting garbaged correctly, which results in ever increasing heap. this leads to the
GC taking more and more time to mark/sweep the heap.
in my case, the HttpSessions were set to expire in 1 week, so over
can you look at the calls trace? and tell us what you see?
having profile tomcat a lot the last 2 years, I personally haven't seen tomcat
increase heap usage with no requests hitting tomcat. If anything, I consistently see
tomcat4 and 5 maintain flat memory usage under constant load. In the
my biased perspective, Borland OptimizeIt is better than JProbe.
the last time I tried to use JProbe to profile Tomcat 4 it was ungodly slow. it's
probably improved since then.
peter
tom ly [EMAIL PROTECTED] wrote:
My team is thinking about getting a profiling tool. Does anybody have any
application is huge AXIS webservices (no typical
servlets here) running remotely on a non GUI (all command line) Linux box and we are
telneting into the box from a windows pc. Can OptimizeIt be running from a non GUI
Linux box, but have statistics shown from a windows box?
Peter Lin wrote:
my
ok, so your server has 2GB of ram or 2000Mb.
the memory usage may increase for several reasons, the
primary one is there's a slow leak in one of your
applications. This may be as simple as several objects
referencing each other. If they happen to result in
circular references, the garbage
by the way, you can view the performance of Tomcat using JMeter.
it will make a request to the status page and then graph the results.
peter
Shapira, Yoav [EMAIL PROTECTED] wrote:
Hi,
Here is a blurp i found searching through google for memory top and
tomcat:
It wouldn't hurt to quote
Randall Svancara schrieb:
I remember you talking about adding that functionality to jmeter. It sounds really
kewl. I need to try it!!
-Original Message-
From: Peter Lin [mailto:[EMAIL PROTECTED]
Sent: Thursday, May 06, 2004 12:13 PM
To: Tomcat Users List
Subject: RE: Question about
I've profiled and tested the new monitor plugin for JMeter. under constant load, the
memory usage of jmeter is flat :)
OptimizeIt really helps achieve flat memory usage under constant load. JMeter has an
old mailer that will send out EMails if a set number of requests fail, so JMeter is
decent article and a good introduction to tomcat5's
clustering.
thanks for posting it.
peter
--- Ralph Einfeldt [EMAIL PROTECTED]
wrote:
A two part article on clustering and loadbalancing
with tomcat:
http://www.onjava.com/lpt/a/4649
http://www.onjava.com/lpt/a/4702
I think the current status servlet provides request count per page and webapp. I could
be wrong. have you looked at the full status page in TC5?
peter
Sankaranarayanan (Ganesh) Ganapathy [EMAIL PROTECTED] wrote:I see that tomcat
provides JMX administration beans however I don't see
tomcat
/application
frameworks to make .NET look better. seems strange.
peter lin
Alex Gibson [EMAIL PROTECTED] wrote:
Hello All,
Here is a image to show how it behaves on file downloads. However, when I
upload a file there is no issue, it uses as much of the bandwidth as
possible.
Alex Gibson
From my own experience, this kind of behavior appears when a session isn't getting
timed out for one reason or another. For example, say you get data from some remote
site using your own Http client libraries that is multi-threaded. If that thread sits
around and the socket it has isn't
methods
are getting called. Usually, that is enough to point towards a culprit.
repeat, and rinse as many times as needed until you've squashed all bugs and leaks.
peter lin
Matt Woodings [EMAIL PROTECTED] wrote:
I just read your post this morning (I am lurking today as I have a few
issues
to log off BOA and not just the section I'm on. Other people might have
different expectations, but that's how I tend to think of single signon.
peter lin
Summers, Bert W. [EMAIL PROTECTED] wrote:
I am using the SingleSignOn class from Tomcat.
It is working good in that I have three webapps
it. that's my
biased perspective :)
peter lin
--- Allistair Crossley [EMAIL PROTECTED]
wrote:
This evening I did the same with JProbe under 360
requests as you describe and this led to me gaining
a 38% speedup.
The main bottleneck I found was some very simple
tags I have were calling out.flush
some of the tomcat developers have blogs, but they don't necessarily talk about tomcat
in their blogs. some do.
there is not unified tomcat blog site, unless you are proposing to create one and let
everyone use it :)
peter
tomcatuser [EMAIL PROTECTED] wrote:
I am talking about a weblog
stress test. If it looks suspicious, I run more tests. so far it's
worked well for me and increases my chances of delivering a solid app.
peter lin
Shapira, Yoav [EMAIL PROTECTED] wrote:
Hi,
Wonder wonder wonder ;) Pick up a profiler (you can get free evals),
find the leak, and post your
why do you want to use a thread to manage authentication? given the requestProcessor
threads are reused, it makes no sense to use the thread for the mapping.
you're better off just authenticating the first time and setting the HttpSession,
rather than look up the thread. I'm probably missing
but it looks like there will be lot
of problems.
Ross
-Original Message-
From: Peter Lin [mailto:[EMAIL PROTECTED]
Sent: Thursday, April 01, 2004 11:21 AM
To: Tomcat Users List
Subject: Re: Threads in Tomcat 5
why do you want to use a thread to manage authentication? given
in or
uploading a file with MultipartRequest.
The applications were originally created 2-3 years ago to work on iPlanet
and now I'm trying move them to Tomcat but it looks like there will be lot
of problems.
Ross
-Original Message-
From: Peter Lin [mailto:[EMAIL PROTECTED]
Sent
information. Each http session is related to an
eRights session.
Ross
-Original Message-
From: Peter Lin [mailto:[EMAIL PROTECTED]
Sent: Thursday, April 01, 2004 12:26 PM
To: Tomcat Users List
Subject: RE: Threads in Tomcat 5
hit send too soon. Another approach which I've used and is common
. In fact in .NET
1.0, disk caching causes heap fragmentation and memory
leaks. If disk cache is kicking in, the performance
issues may be the result of heap fragmentation, which
kills GC performance.
peter lin
--- sp k [EMAIL PROTECTED] wrote:
We have a Tomcat site which becomes terribly slow
You're going to have to qualify your definition of scalability before anyone can
provide useful information.
scalability in terms of concurrent users?
requests per second?
average response time?
cluster size?
concurrent connections?
without a point of reference, scalability means very little
101 - 200 of 462 matches
Mail list logo