We associate the SLSB to the custom-rmi-invoker in our jboss.xml file:
| session
| ejb-nameOptimizerPilot/ejb-name
| jndi-nameOptimizerPilotBean/jndi-name
| local-jndi-nameOptimizerPilotLocal/local-jndi-name
| configuration-nameSniffer
The custom-rmi-invoker has not been associated with any container (certainly
not the Sniffer Stateless SessionBean shown here. See the
invoker-proxy-binding-name in the jboss_4_0.dtd:
| !ELEMENT container-configuration (container-name , call-logging?,
| invoker-proxy-binding-name?,
I have done some more testing on this.
I turned on the TRACE for org.jboss.invocation and found some strange results.
I see the TRACE logging when JBoss is starting up but when my Servlet is called
and does the service lookup and subsequent call on the SLSB I do not see any
TRACE code for
I think I am having a similar problem to you. I have servlet that looks up the
SLSB on HA-JNDI. The behaviour I see is that the only SLSB that gets called is
the one living on the same JBoss instance as the servlet.
My guess is that the reason for this is the optimization that JBoss does when
Scott,
Thanks for the response. I looked at the comments for JBAS-1442 but I am still
a little confused. I am using jboss4.0.2. The comments by Adrian Brock on 22
Feb 05 seem to indicate that this problem was fixed, but I am still seeing the
incorrect behavior. To get the correct behavior
http://jira.jboss.com/jira/browse/JBAS-1442
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=3904656#3904656
Reply to the post :
http://www.jboss.com/index.html?module=bbop=postingmode=replyp=3904656
---
Certainly disabling the call by reference invoker will fix the issue, but this
should be working with the default invoker interceptor in 4.0.2. There is a
org.jboss.test.cluster.test.CrossServerCallsUnitTestCase that validates
repeated calls to a SLSB from a server in a cluster does in fact
[EMAIL PROTECTED] wrote : Have you tried to test just SLSB clustering first
without MDB? Check out the example there.
| http://wiki.jboss.org/wiki/Wiki.jsp?page=JBossHA
|
I've done this - instead of calling the SLSB from the MDB (in my basic test
example I linked to earlier), I'm calling
OK, I know it is only being run on one node at a time, but what I'm suggesting
is that if you change the attribute for ReceiversImpl to use the
ReceiversImplLinkedList for the queue or ConnectionManager, you can make each
message get handled on a different node, each node in turn.
In my own
Just so we're clear, this round-robin MDB handling is running using
HAJMA/HASingleton, so fail-over works too.
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=3903469#3903469
Reply to the post :
You don't make it clear if the MDB is being run on the different nodes in
round-robin fashion. Is it? (Have you specified one of the List-based
ReceiversImpl implementations that makes round-robin work?)
View the original post :
The MDB is run on just one node at a time, using HAJMA/HASingleton
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=3903311#3903311
Reply to the post :
http://www.jboss.com/index.html?module=bbop=postingmode=replyp=3903311
toadie wrote : joff wrote : (via a loadbalancing device in front of them)
| |
| can't you use the same device to round-robin the request for the ejb client
stub for you? especially since your bean is stateless
I'm afraid not, the loadbalancer is only there to handle requests from the
joff wrote : (via a loadbalancing device in front of them)
|
can't you use the same device to round-robin the request for the ejb client
stub for you? especially since your bean is stateless
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=3902510#3902510
Have you tried to test just SLSB clustering first without MDB? Check out the
example there.
http://wiki.jboss.org/wiki/Wiki.jsp?page=JBossHA
I suspect it is just some setup problem. But if you are really pressed for time
for production, you should consider subsribe to JBoss support.
Forum post
Hi - the next version of my app is about to go into production, but this issue
is still unresolved, and it runs jobs that can take up to a couple of hours to
run - being able to balance the load of these jobs across my app servers will
be a major help. can anyone shed any light as to why
i've also just added in
configuration-nameClustered Stateless SessionBean/configuration-name
for my EJB DDs, but still no joy :-/
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=3893301#3893301
Reply to the post :
*bump*
Is anyone able to help me here? I'm no closer to finding the solution...
View the original post :
http://www.jboss.org/index.html?module=bbop=viewtopicp=3887605#3887605
Reply to the post :
http://www.jboss.org/index.html?module=bbop=postingmode=replyp=3887605
Hi Ben
At the URL below is an EAR containing a minimal test case for my problem. It's
basically a servlet for triggering the operation, an MDB to receive the JMS
message, which calls a SLSB ten times
Souce is also included
http://joff.googlehax.com/LBTest.zip
PS. You will either need to add
If you still have a problem, can you attach a sample client and jar file to
troubleshoot it?
-Ben
View the original post :
http://www.jboss.org/index.html?module=bbop=viewtopicp=3884116#3884116
Reply to the post :
http://www.jboss.org/index.html?module=bbop=postingmode=replyp=3884116
Oops, hit 'submit' instead of 'preview' anyway.. to continue:
the jboss.xml for the SLSB is:
session
| ejb-nameTransactionConductor/ejb-name
| jndi-nameTransactionConductor/jndi-name
| local-jndi-nameTransactionConductorLocal/local-jndi-name
|
21 matches
Mail list logo