As I understand it, the failover should be handled by
the external load balancer.
Why shouldn't it work with uneven number of tomcats ?
You should just have to define an odd number of workers
in each mod_jk.
> -Original Message-
> From: Peter Anning [mailto:[EMAIL PROTECTED]
> Sent: T
See below:
> -Original Message-
> From: Peter Anning [mailto:[EMAIL PROTECTED]
> Sent: Thursday, August 28, 2003 10:44 PM
> To: Tomcat Users List
> Subject: RE: mod_jk load balancing with multiple apache servers
>
> If Apache yet it's asociated tomcat is down is up then the
> loadbalance
In my experience this is not a good recommendation:
- -server is less stable than -client in all JDK's that I tried,
and this has been confirmed by several list members.
- -server won't help much on out of memory errors. The gc is
behaving differently, but it can't free more objects, they
httpd instance doesn't "autodiscover" the other tomcat,
you have to configure it.
I would suggest the following:
- define distinct propertie fieles for each mod_jk
- define 2 workers for each mod_jk
- each worker points to one tomcat
- set the lbFactor for worker1 to 0 for one mod_jk
- set the lb
We prefer to use CATALINA_BASE and CATALINA_HOME.
This decouples the application best from the tomcat
installation and you still can set all set all
parameters individual:
http://archives.real-time.com/pipermail/tomcat-users/2002-August/074367.html
(In short the trick is to do the settings in
In the archive you can find this:
http://archives.real-time.com/pipermail/tomcat-users/2002-August/074367.html
> -Original Message-
> From: Pitre, Russell [mailto:[EMAIL PROTECTED]
> Sent: Thursday, September 04, 2003 11:01 PM
> To: Tomcat Users List
> Subject: RE: Tomcat and multiple pro
There are two options:
- Use sticky session.
In this scenario you don't need session sharing to do
load balancing, but it would make the system more fail
safe. (If one tomcat gets out of service, sticky sessions
without session sharing loose all data from the current
sessions on this
It depends on the configuration.
I guess that you are talking about apache and tomcat
on different maschines (server is a bit ambigous).
In that scenario you either have to setup mod_jk to serve
the static files, or you have to mount (on the os level)
the document root from server B to the docu
I think the next step can be one of the following:
- Work with 2 workers in each apache
- connect httpd0 with tc1a
- connect httpd1 with tc0a
- remove tc0b and tc1b
- Work with 4 workers:
- add 2 workers in httpd0 that point to tc1a and tc1b
- add 2 workers in httpd1 that point to tc0a
mx is limiting the heap of a java process.
In other words:
It limits the total heap for one tomcat instance.
The total memory can be much higher:
- stacksize per thread
Some vm's let you set these values (-Xss).
If you have some hundred of threads that can make up some memory.
- static memo
Sorry can't give a recommendation.
We had different results even for (nearly) the same
software for different sites. For one site the IBM JDK
worked best, for the other Sun's.
Any given result may be wothless with the next version
of either jdk. (Even with just changes in the minor
revision)
- What is the compiler complaining about ?
BTW: It's not legal to cast a String[] to an ArrayList.
This should at least produce a runtime error.
You have to explicitly convert the String[] to an ArrayList.
ArrayList importList = new ArrayList(Arrays.asList(request.getAttribute("importList"));
>
Not with JSP. Only with javascript.
(Which isn't covered in this list :} )
> -Original Message-
> From: Anson Zeall [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, September 17, 2003 12:31 PM
> To: Tomcat Users List
> Subject: Using drop down boxes in JSP
>
> I want to know, what is the best
Reloading the webapp may help. AFAIK it reloads all classes
that are bound to the webapp classsloader and thus may free
some memory that is hold by those classes. (Not that I think
that it's a good idea to rely on that)
> -Original Message-
> From: Shapira, Yoav [mailto:[EMAIL PROTECTED]
So it's the vm that crashes.
Some recommendations (quite unsorted):
- There should be a core file from the vm somwhere
With that you may find out where the error happened.
- Find out if there are recommended patches for your os and your vm
- If you can't find such patches, you may have to upgra
It looks to me like something screwed up your system.
The main cause is not within tomcat.
The important part of the traceback is this:
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.jav
a:39)
I think it's not the right time to think about optimizing.
If tomcat isn't even starting, I don't think that the
problem is a 'normal' memory leak especially if the server
has run before. I think that something has changed through
the reboot.
> -Original Message-
> From: Denise Mangano
May setting these values will help:
ulimit -s 2048
export LD_ASSUME_KERNEL=2.2.5
Either put them in the environment of the user that starts tomcat
or put it in the startscript of tomcat.
BTW: Have you looked what processes are running and
how the memory usage is, before you try to start tomcat ?
Yes, that is my assumption.
The processlist looks uncritical.
What about the memory usage ?
(I'm not familiar with red hat, 'top' or 'cat /proc/meminfo'
should work)
> -Original Message-
> From: Denise Mangano [mailto:[EMAIL PROTECTED]
> Sent: Friday, September 19, 2003 4:22 PM
> To: To
Yes, that looks fine to me.
I think now you have to consider the other tip like
- run fsck
- Reinstall java
- Reinstall tomcat (but I can't believe that this really helps)
if that doesn't help it getting hard, as there is some knowledge
required. (I would try to play with strace to find out what
Some other recommendations:
- Try to diable as much as possible in server.xml and web.xml.
- Try to run tomcat standalone.
Just to see if it starts. If it starts, try to find out
which feature caused the error.
-
To unsubscribe
Install the second instance in a new directory tree.
Diasble all unneeded ports.
Set the rest of the port to unique values.
Start this instance.
As first step I wouldn't add any of your applications.
Just get the plain tomcat to run. (In the first run with
the least possible set of features, thos
If ps doesn't show any, there aren't any.
What make you think there are ?
> -Original Message-
> From: Denise Mangano [mailto:[EMAIL PROTECTED]
> Sent: Monday, September 22, 2003 5:50 PM
> To: Tomcat Users List
> Subject: RE: Cannot run Tomcat as new user [Was: Urgent! Cannot start
> Tomc
May be you should run
ulimit -a # sh/bash as shell
limit # tcsh as shell
for the root user and the user you wann't to start tomcat with.
Maybe that shows something interesting.
> -Original Message-
> From: Denise Mangano [mailto:[EMAIL PROTECTED]
> Sent: Monday, September 22, 2003
d to
> believe that there is no solution.
>
> Thanks,
> Denise
>
> -Original Message-
> From: Ralph Einfeldt [mailto:[EMAIL PROTECTED]
> Sent: Monday, September 22, 2003 12:10 PM
> To: Tomcat Users List
> Subject: RE: Cannot run Tomcat as new user
>
>
&g
More a OS parameter.
(As otherwise tc wouln't run as other user)
E.G.:
max user processes (-u) 20
max memory size (kbytes, -m) 5000
These are quite low. We have 256 and unlimited.
20 processes per user and especially 5MB memory is
not enough to run tomcat. (I'm not shure if these
values are ig
I'm quite shure that the smallest possible tomcat
installation won't run with 5 MB memory. So I think
that is nothing that has anything to do with your
application. (The default installation of tomcat
on our machine needs 20 MB)
Although I don't know what happened, I could bet
that these setting
CVS has a gui interface. (Even more than one)
http://www.wincvs.org/
http://www-csr.bessy.de/control/Soft/tkCVS/
http://www.tortoisecvs.org/
For a more complete overview see:
http://www.cvshome.org/dev/addons.html
> -Original Message-
> From: Pitre, Russell [mailto:[EMAIL PROTECTED]
> S
Tomcat uses a ThreadPool so Threads are recycled.
If you store something in a ThreadLocal it it your
resposibility to clear the stored values at the end
of the use. So if you store request variables you
have to clear them at the end of the request (or
before storing them at the beginning of th
Not true.
You also can use xinetd to redirect requests to one port
to another port where tomcat is running:
http://www-106.ibm.com/developerworks/linux/library/l-secjav.html
> -Original Message-
> From: QM [mailto:[EMAIL PROTECTED]
> Sent: Friday, April 02, 2004 12:43 AM
> To: Tomcat Use
That isn't easy to answer, that depends on your infrastructure
and on your security requirements.
xinetd isn't required. You can achive the same (in the context
of your question) with iptables (Which is more efficient) or
with commons-daemon:
http://jakarta.apache.org/commons/daemon/
> -Ori
That information is stored in the session.
So your programm has to include the session id
that was created by tomcat in the requests (either
as cookie or as query parameter)
> -Original Message-
> From: Malcolm Warren [mailto:[EMAIL PROTECTED]
> Sent: Friday, April 02, 2004 10:12 AM
> T
That is not a tomcat but an jvm issue.
You are confusing me.
You say the problem happens in
1.4.2_04 Client VM (build 1.4.2_04-b05, mixed mode)
The dump says the problem happens in
java HotSpot(TM) Client VM (1.4.2-b28 mixed mode)
You say this one works.
Can you please verify which one wor
It's technology by intel:
http://www.intel.com/products/ht/hyperthreading.htm
> -Original Message-
> From: Zsolt Koppany [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, April 14, 2004 8:41 AM
> To: Tomcat Users List
> Subject: RE: IBM Java VS SUN Java
>
>
> Hi,
>
> sorry for the question
So it's most likely the kernel or an os library for
Red Hat Enterprise Linux ES release 2.1 that is
used by java that causes the problem.
As I don't have any experience with Red Hat
I can't provide any help in this environment.
- Are there updates to the kernel or the core libraries ?
> -
Other potential causes:
- Programming errors of programmers that don't understand
the impact of multi threaded programms in a servlet
container on the scoping of variables.
- Caching of proxies between the browser and the server.
Make shure that you set the headers of the response.
http:
Some members of this list have reported that the load
balancing is broken as the instances of mod_jk[2]
don't know the load of each other. So mod_jk[2] will
balance to some extend but not as good as it could/should.
> -Original Message-
> From: Thomas D. Zeimet [mailto:[EMAIL PROTECTED]
Did you set the jvmRoute in the Engine tag in server.xml.
The value must match the name of the workers.
(I'm not shure if it's the name or the attribute 'tomcatid')
> -Original Message-
> From: ian [mailto:[EMAIL PROTECTED]
> Sent: Thursday, April 15, 2004 9:34 AM
> To: 'Tomcat Users List
- You don't have to implement your own url re-writing
- The time is long gong when the session id was a parameter,
it's now before the ?
Just throw away your url rewiting and use rsponse.encodeUrl()
instead.
> -Original Message-
> From: Malcolm Warren [mailto:[EMAIL PROTECTED]
> Sent:
Do you want to use other contexts beside those two ?
If not, you can use one context with an empty path
so that fred and fred2 are just subdirectories of
one context.
> -Original Message-
> From: Ben Janes [mailto:[EMAIL PROTECTED]
> Sent: Thursday, April 15, 2004 4:32 PM
> To: Tomcat
Is it possible that mails from the list are blocked
by the new host ? (Blacklisted or identified as spam)
> -Original Message-
> From: John Trollinger [mailto:[EMAIL PROTECTED]
> Sent: Thursday, April 15, 2004 4:36 PM
> To: [EMAIL PROTECTED]
> Subject: List problems
>
>
> I just changed
The main point was not, how many you want to use,
but if you want a real contexts beside the others
that have their own context element.
Trying to rephrase the question:
Do you want all paths handled by one context ?
If yes you can follow my suggestion.
> -Original Message-
> From: Ben
>From your description it is not shure that you have a
memory leak at all.
The vm is not returning free memory to the os. So the
memory as seen by the os will alway be the maximum value
that the jvm ever needed during the runtime.
The other option that explains your observertion is that
you tes
d balancer), so no traffic is going to
> it at all the
> memory doesn't go down.
>
>
> Chanan Braunstein
> Knovel Corp.
> Web Development Manager
> 607-773-1840 x672
> http://www.knovel.com
>
>
> -Original Message-
> From: Ralph Einfeld
Which vm and os do you use ?
That behaviour is jvm and os dependend. (This is the first
time I hear of an implementation that returns memory to the
os, although I knew that it could be done)
> -Original Message-
> From: Allistair Crossley [mailto:[EMAIL PROTECTED]
> Sent: Thursday, Apr
- Tomcat is a servlet and jsp engine.
- Tomcat can be used as a stand alone webserver
- Apache is a webserver (with different features than tomcat,
if it is better depends on the requirements)
- Apache can be integrated with tomcat by mod_jk[2]
(So Apache replaces tomcats own http stack)
-
I think it will work if you change docBase to "webdirs"
> -Original Message-
> From: Ben Janes [mailto:[EMAIL PROTECTED]
> Sent: Tuesday, April 20, 2004 7:56 AM
> To: Tomcat Users List
> Subject: RE: Multiple Paths in one Context
>
>
> So now I have context definition
>
I don't know anything about axis.
The message indicates for me that axis is reading from a http server.
The server answers (with 404).
The question is, which http server is it talking to ? May be
it's just talking with the wrong http server.
Are you using tomcat standalone or behind apache/iis/
Just to make shure that I understood everthing:
- The problem exists in the development setup
- Tomcat is running standalone in the development setup.
I just can think of the following causes:
- The servlet in tomcat is talking to the same tomcat
instance.
- The servlet in tomcat is talking
> - The servlet in tomcat is talking to the same tomcat
> instance. [-?- The servlet running in tomcat will get it's
> data from a physically different server -?-]
Unless the error message is heavily missleading the servlet/axis
is talking to a webserver. If it isn't the iis on server B it
You just have to remove the token no matter what happens.
try {
// store token
// Do the processing and create output
} catch(SomeException ex) { // Optional
} finally {
// remove token
}
> -Original Message-
> From: Antony [mailto:[EMAIL PROTECTED]
> Sent: Thursday, April 22
A two part article on clustering and loadbalancing with tomcat:
http://www.onjava.com/lpt/a/4649
http://www.onjava.com/lpt/a/4702
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
AFAIK the session id has a random part so you need
either quite some luck to guess the number or you
have to find a bug in the random generator.
> -Original Message-
> From: Bill Gorr [mailto:[EMAIL PROTECTED]
> Sent: Friday, April 23, 2004 5:35 PM
> To: [EMAIL PROTECTED]
> Subject: sess
I would go for 2.
As I'm not using IIS, I'm not shure if there are specific
hurdles in this environment. Tomcat stand alone and tomcat
behind apache are quite easy to setup with multiple
instances.
If all tomcate have the same versions, you just have to
install tomcat once and setup two instan
Guessing that you are talking about starting instead of stopping...
is for servlets not for web applications.
If you include this tag in a servlet definition, the servlet will
be started at startup time. The value for the tag defines the order
of the startup.
> -Original Message-
> Fr
It's not always easy easy to find the right list for
a question. The opinions about what is right or wrong
vary with the members of the list. (In my opinion
your question is right in this list, although I can
understand that others have different opinions in this
case)
I have seen questions tha
As I understand it, the RemoteAddrValve works on the address
of the client (browser) not of the webserver.
I'm not using mod_jk. Back in the old days of mod_jserv
there was a property ApJServSecretKey to protect a jserv
instance from unauthorized access. I'm not shure if this
feature survived i
???
He says that the error was the same when using submit buttons
and you say he should use it.
AFAIK your explanation is wrong.
If you submit a form with javascript (even if triggered by a link)
it uses (should use ?) the method type that is given in the form.
> -Original Message-
> Fr
What is your environment ? (os, tomcat, apache/iis, jdk)
Find out if the request that reaches your server is a POST:
- Look for the request in the access log.
- use an ethernet sniffer to see what is coming.
after this test you know if you have to look at the server
or the client side.
> -O
- stability
No more to say than Yoav.
- performance
You may observe better performance with several vms on one server:
- Depending on the quality of the thread implementation
in your java vm and the os.
- Depending on the quality of the gc implementation you may
gain per
I don't think that the problem is caused by the tomcat side.
The output looks pretty much like a memory or a disk dump.
You normally can't access memory outside of the java process
from within tomcat. (Native libraries excluded)
So I think it's mod_jk, apache or the jvm that cause this problem
I don't think that a debugger will help much in this case.
A profiler might help, if you can reproduce the error in a
test environment. (see below)
Some ideas:
- Use the access log to try to reproduce the error in a
test environment. (AFAIK apache jmeter has the ability
to create a test
public void reset() {
oEndTime = 0;
}
public void timeout() {
oStartingThread.interrupt();
}
}
> -Original Message-
> From: Ralph Einfeldt
> Sent: Friday, May 07, 2004 7:12 PM
> To: Tomcat Users List
> Subject: RE: How to limit time for Connector
In some of the windows versions the number of concurrent
inbound tcp connections is limited.
http://support.microsoft.com/?kbid=122920
http://support.microsoft.com/default.aspx?scid=kb;en-us;328459
I'm not shure if this is still true for xp and 2003.
(But I think it is)
-Original Message--
Has your mail address changed or do you now use a different
address than in your subscription ?
Look at the return path of the mails you receive from the list
and you get:
Return-Path: tomcat-user-return-26788-=@jakarta.apache.org
If that doesn't match roy=panix.com
use this address to unsubscr
That statement from the site is wrong in at least one case:
page import
May be it's wrong in the other cases too.
AFAIK the spec doesn't mention such restriction.
(At least I couldn't find it)
> -Original Message-
> From: Jason Irwin [mailto:[EMAIL PROTECTED]
> Sent: Wednesday,
Who wrote what of the Servlet:
http://cvs.apache.org/viewcvs.cgi/jakarta-tomcat-catalina/catalina/src/share/org/apache/catalina/servlets/CGIServlet.java?annotate=1.19
> -Original Message-
> From: Alex [mailto:[EMAIL PROTECTED]
> Sent: Friday, May 21, 2004 2:29 PM
> To: Tomcat Users List
I'm not shure about the response code that is returned
in both cases.
It's possible that tomcat returns a different response code
for a html and a jsp. (Like 200 for html because the html
error page was found and some other value for the jsp because
the jsp explicitly sets an response code.)
One possible cause:
The component that store your object find the class through
a different classloader than the component that tries to
retrieve the object.
Where do you store the class/jar for the stored object ?
> -Original Message-
> From: Rui Zhang [mailto:[EMAIL PROTECTED]
> Se
Is the other web server also tomcat or something else ?
> -Original Message-
> From: Gunnar Pörschke [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, June 09, 2004 4:09 PM
> To: 'Tomcat Users List'
> Subject: AW: AW: IP Adresses
>
>
> Thanks, this time it helps. But Tomcat still blocking all
Although it's quite off topic:
In /etc/mysql/my.cnf comment out this line:
skip-networking
restart mysql
> -Original Message-
> From: ilasno [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, June 09, 2004 5:13 PM
> To: Tomcat Users List
> Subject: Re: unable to connect to mysql db on localh
AFAIK there is no way to disable this by configuration.
Just an idea:
Implement a servlet filter and a custom ResponseWrapper.
The ResonseWrapper can have it's own encodeURL that returns
the url unmodified.
Map that filter to every request.
As I havn't tried it I'm not 100% shure...
> -O
there are two things to consider:
- the time that is spent in arguments of the call.
- the time that is spent in the call
If you have something like
debug("abc")
there is hardly any time spent constructing
the argument. But if you have
debug("abc" + someNestedObject.toString());
or
debug(tr
I would expect that the Appenders and Formatters only come
into play if debug is enabled. Am I wrong ?
> -Original Message-
> From: Shapira, Yoav [mailto:[EMAIL PROTECTED]
> Sent: Thursday, December 04, 2003 3:34 PM
> To: Tomcat Users List
> Subject: RE: How to measue performance? Log4j o
jServ is just a servlet engine (JSDK 2.0).
Tomcat is a servlet engine (JSDK 2.2/2.3/2.4),
a jsp engine and a webserver and contains much
more additional functionality (DB Connection Pool,
Realms, Manager, )
> -Original Message-
> From: Jan Behrens [mailto:[EMAIL PROTECTED]
> Sent: Sa
If it really is not allowed, then the handling of several proxies
and/or browsers is severly broken. We had several problems with
caching until we forced the reload by appending a unique query
parameter in each link.
> -Original Message-
> From: Tim Funk [mailto:[EMAIL PROTECTED]
> Sent
First of all keep in mind that it aren't processes, they are
threads. As the threads share the same memory they do not
use much additionals memory (some memory for the thread local
objects and the thread stack and so on)
I can't explain all your numbers.
- Even the simplest java program has mor
What about using one centralized servlet that parses
req.getPathInfo(), sets the language as request attribute
and forwards to the real servlet(s) ?
> -Original Message-
> From: Marten Lehmann [mailto:[EMAIL PROTECTED]
> Sent: Tuesday, December 09, 2003 11:26 PM
> To: [EMAIL PROTECTED]
>
Sorry not much time, try searching for request dispatcher and forward
> -Original Message-
> From: Marten Lehmann [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, December 10, 2003 4:46 PM
> To: Tomcat Users List
> Subject: Re: servlet-mapping like mod_rewrite?
>
>
> Hello,
>
> > What about
Although my solaris days are long gone I would guess that
this aren't 'real files' but memory mapped files.
Are you shure that there are several handles on the same file,
or is the same handle just reuses several times ?
> -Original Message-
> From: Hollerman Geralyn M [mailto:[EMAIL PR
IIS bind per default to all interfaces.
http://support.microsoft.com:/support/kb/articles/Q238/1/31.ASP
> -Original Message-
> From: Ryan Golhar [mailto:[EMAIL PROTECTED]
> Sent: Thursday, December 11, 2003 4:02 AM
> To: 'Tomcat Users List'
> Subject: RE: How to bind to multiple IP addre
You are mixing several differrent things:
- J2EE is a spec that consists of several technologies.
Most of these technologies can stand for them selves
and be used without the others. (Some of them are
related or even dependend to some extend)
- JBOSS is an implementation of that spec.
It
JBoss and J2EE SDK rely on the same tools for some
technologies. Others technologies are implemented
from scratch. From the J2EE perspective JBoss and
J2EE SDK are competitors that have different scopes.
(J2EE SDK the is not intended as a tool for production)
The heavy part (like EJB) is imple
Sun want's to sell their Sun ONE software, so the
licence for the free J2EE SDK is restricted. (For
internal use and development)
(But this is getting quite off topic for this list)
> -Original Message-
> From: Michael Coughlan [mailto:[EMAIL PROTECTED]
> Sent: Monday, December 22, 2003
You can, but in the most cases you shouldn't.
It is not recommended because you may consume much
memory this way. Per default a session times out
after 30 minutes of in activity.
If the everage session lasts 35 minutes and you get 2
new sessions per Minute you have already 70 concurrent
sessi
- Nobody (the spec included) prevents you from creating threads
You just have to follow some rules.
- Besides creating threads I prefer to use servlets/jsp's
that are triggered from the outside to do regular jobs.
(Using cron and wget to trigger the action)
This way you don't have to deal
The cronjob is requesting an url with wget
and triggers tis way a servlet.
That's easy and sufficient to do regular jobs
like sending newsletters, cleaning up files,
that don't require any interaction)
> -Original Message-
> From: Nikola Milutinovic [mailto:[EMAIL PROTECTED]
> Sent: Mon
The servlet has access to all information inside the container/context
that may be needed to do the job. For an external job this can be much
harder. (Iterate over all sessions to do something with them, access
attributes with application scope, access attributes from server.xml/
web.xml) It's
There is an unneeded .
As I understand the error message you have to use
either or .
(Not shure if this is the problem, as I havn't used filters by now)
> -Original Message-
> From: Patrick Scheuerer [mailto:[EMAIL PROTECTED]
> Sent: Tuesday, January 06, 2004 5:32 PM
> To: Tomcat Users
Tomcat is not clustered. It just allows session
replication between tomcat nodes.
AFAIK every thing else is unique to each instance.
You have to deploy to each instance, and you have
unique pools in each instance that don't know
anything about each other.
(Not quite shure about the pool, have
That indicates either
- that your servername and/or port in the connection string is not
correct.
- server and port are not reachable from the machine that runs
tomcat (e.g. the port is blocked by a firewall).
- the database isn't running.
> -Original Message-
> From: [EMAIL PROTECT
May be you should have a look at the following:
- the package statements in the generated source files
- the file structure of the generate class files
> -Original Message-
> From: Jay Glanville [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, January 07, 2004 2:20 PM
> To: 'Tomcat Users List
The value CATALINA_HOME doesn't sound valid.
CATALINA_HOME should point to the directory that
contains typically something like:
bin
webapps
lib
server
conf
common
logs
work
classes
> -Original Message-
> From: Jason Roscoe [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, January 07, 2004 3:
You either have to download the binary version of tomcat
from http://jakarta.apache.org/site/binindex.cgi
or to create a binary from the sources:
http://jakarta.apache.org/tomcat/tomcat-5.0-doc/building.html
More details installing the binary download can be found in
http://cvs.apache.org/viewcv
3) Request contains no session id (URL/Cookie)
4) Request contains invalid session id (URL/Cookie)
Have a look at
- isRequestedSessionIdFromCookie
- isRequestedSessionIdFromUrl
- isRequestedSessionIdValid
- getRequestedSessionId
in HttpServletRequest to find out if the request
contains a session
Are you using tomcat standalone or behind apache/iis ?
In the later case there may be some buffering in
the webserver or mod_jk that can't be controlled
from within tomcat.
> -Original Message-
> From: Vitor Buitoni [mailto:[EMAIL PROTECTED]
> Sent: Thursday, January 08, 2004 1:37 PM
>
See below:
> -Original Message-
> From: T K [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, January 07, 2004 9:03 PM
> To: [EMAIL PROTECTED]
> Subject: weird tomcat questions
>
>
> * How does scoping work in Tomcat?
What kind of scoping do you mean ?
- attribute scoping a la servlet spec
One thing i forgot in my previous reply:
To implement singletons under tomcat requires more
knowhow as a typical stand alone applications.
Tomcat has several classloaders and there can a copy
of the same class in different classloaders. This way
the simple approach of using static class vars to
If that takes to long, you can limit the search,
if you have a table with an indexed column
where you know that there is certain id
SELECT 1 FROM WHERE =
(We always have such tables)
> --- Allistair Crossley <[EMAIL PROTECTED]>
> wrote:
> > i think you could use anything .. maybe
> >
> > SE
One additinal remark:
In the long term having one instance with multiple
webapps can easily become a maintainance and/or qs
nightmare.
With each webapp in a dedicated instance it's much
easier to change versions of libraries or even the vm.
Just imagine what happens if one of 100 applications
101 - 200 of 1290 matches
Mail list logo
