So what I understand is that if I have only 1 app deployed in the cluster, or
if I have multiple apps that dont rely on sharing sessionIds, then turning off
the recycling will guarantee a new sessionId is generated for each new session.
Is that accurate?
Thanks.
View the original post :
Brian,
It was reported to me in October, then again today. I log the sessionId when a
user logs into my app, and today I could see that the sessionId
hGbGpyqSV2CPfJKGZi0KGg**.node1
was given to one user at 1:49PM and then to another user at 4:43PM. The first
user complained that at 4:45PM
I am seeing this problem as well. I am using JBoss 4.2.3 behind mod_jk for
load balancing. I have 4 nodes in the cluster. I use REPL_ASYNC for session
duplication.
I do have distributable/ in my web.xml
Here is my jboss-web.xml
jboss-web
|replication-config
|
I am having a weird, sporadic and impossible to duplicate problem. If this is
not the correct forum, I apologize and would appreciate a shove in the right
direction.
My configuration is 4 JBoss nodes clustered and running behind the Apache
mod_jk load balancer.
Server Version: Apache
| JK
Last night I reverted back to mod_jk 1.2.21. I have stickysession working
correctly and load balancing working correctly. I no longer have any error
messages regarding the session replication in the jboss log.
This leads me to believe the issue we are all having is with mod_proxy and not
Mr Stansberry,
We're using sticky sessions, we understand how to configure them and as far as
I know they are working correctly, but we are still seeing the Possible
concurrency problem message.
To elaborate a little on our configuration, we use jBoss4.0.5GA with Apache
2.2.4 on the front end
1) I am moving from mod_proxy_ajp back to mod_jk tonight so after than I can be
more sure of sticky sessions working because I know they work for sure in
mod_jk. I know they are not working correctly now, although I have it setup
just like the Wiki explains
This i certainly not a JBoss clustering issue...but...the last time I used
oracle in that mode, you had to use the Type 2 Driver. Hope this helps.
Think of it this way. One of these has to be true.
1)Type 4- The IP of the database never changes even after Oracle's failover
occurs.
2)Type 2 -
What is your environment? JDK/JRE, JBoss, OS?
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=4018609#4018609
Reply to the post :
http://www.jboss.com/index.html?module=bbop=postingmode=replyp=4018609
___
jboss-user
The only difference is that I am using 1.4.2_13 with Linux. I dont know that
much about this, just what I did to fix it. You should check if you have the
Unified class loader setup correctly and then examine the jars in your
server/lib and your EAR to make sure all your classes are covered.
Sree,
You were able to accomplish in a few days what took me and others several
weeks. You are very welcome. Please support the open source forums by
becoming more familiar with the clustering process. Then, monitor this group
and help others. I'm glad you have everything working and I
1) For now, create the directory C:/Program Files/Apache Software
Foundation/Apache2.2/htdocs/TestApp and add a file in there called index.html.
Edit the file to look like this.
| html
| body
| ModJK should not allow me to see this once configured properly.
| /body
| /html
|
2)
Try this:
| # Define list of workers that will be used
| # for mapping requests
| # The configuration directives are valid
| # for the mod_jk version 1.2.18 and later
|
| # Define Node1
| # modify the host as your host IP or DNS name.
| worker.node1.port=8009
|
You need this in your web.xml. Refer to the DTD for the correct order.
| !-- Indicated the app is coded for a clustered environment --
| distributable/
Your JkMount point shoud match the WAR Module Deployment Descriptor in your
EAR's application.xml file.
module
| web
|
I had this issue the other day. It has do do with machines with multiple
NICs/IPAddresses.
Edit the following files:
$JBOSS_HOME/server/yourserver/deploy/cluster-service.xml
$JBOSS_HOME/server/yourserver/deploy/tc5-cluster.sar/META-INF/jboss-service.xml
Look for this comment where the
Set the DocumentRoot to any valid Directory on your Machine. It wont matter
since you are proxying all requests to JBoss. set it to c:\temp.
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=4016497#4016497
Reply to the post :
Install Apache and configure the nodes as described in the wiki for using
mod_jk1.2 with JBoss. The .so file _IS_ correct. I installed apache on my
windows machine to verify that.
.so files are not made like .jar files. JAR files are just zip files with a
.jar extension. .so files are
yes.
anonymous wrote : Install Apache and configure the nodes as described in the
wiki for using mod_jk1.2 with JBoss. The .so file _IS_ correct. I installed
apache on my windows machine to verify that.
|
View the original post :
With apache in place, you _never_ go to the JBoss port again. Always go
through apache. It monitors the server states and directs traffic accordingly.
Apache knows server 1 is dead, so it sends all traffic to server 2. If both
server 1 and server 2 are alive, it load balances reuqests
Apache is configured with the file $APACHE_HOME/conf/httpd.conf
I suggest you read up on Apache server configuration as there is very good
documentation on it.
In that file you should have a listen port setting. Yours is obviously 80
based on your feedback. Here is where I think you have
Sree,
My configuration was with Linux rather than windows. I dont know exactly how
mod_jk gets added into apache as a module, but I am guessing that it is a .dll
file rather than a .so file.
I could be wrong, but you are on your own now. Sorry.
View the original post :
Ok Sree, looks like you have successfully clustered your jboss servers. Now,
you need a mechanism to monitor the state of each server and then direct
traffic accordingly. This explains it all:
http://www.jboss.org/wiki/Wiki.jsp?page=UsingMod_jk1.2WithJBoss
View the original post :
Sree,
I have a couple of recommendations for you.
1) Turn your logging level down to INFO instead of DEBUG. If there is an
ERROR, we can try DEBUG if we need to. Go to server/all/conf and edit your
log4j.xml and modify your CONSOLE appender like this:
appender name=CONSOLE
You need to do a few things.
1) Check your log files when the server starts to make sure the instances see
each other and join the cluster.
2) Try deploying you WAR file in the farm directory of one server and if #1
above is working, then your WAR file will automatically get copied over the
Solution is to add the bind_addr= to the UDP protocol for the tc5 and jboss
cluster service files. Issue is that there are 2 NICs on the servers. Once I
added tha bind_addr everything works great.
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=4011395#4011395
I have 2 physical servers running the same configuration
| JBoss 4.0.5
| J2SDK1.4.2_13
| OpenSuse 10.2
|
I have clustered an application on the 2 physical servers. Both servers
startup fine, but when reviewing the logs I see the following:
| 2007-02-03 13:09:24,507 INFO
26 matches
Mail list logo
