[JBoss Messaging] - Re: long live consumer stops working
by bodrin
anonymous wrote :
| I would suggest you to remove those debug messages... and maybe verify your timeout configurations. Since you are using ServiceBindingManager to a clustered configuration, maybe you didn't update it properly. (Note that there are 4 places you need to replace it on the ServiceBinding).
|
Ok, I suppose you mean 4 places in case of a cluster with 4 nodes, right?
Because my cluster has 2 nodes and I have already updated this 2 places in $JBOSS_HOME/docs/examples/binding-manager/sample-bindings.xml with the attributes from JBOSS_CONFIG/deploy/jboss-messaging.sar/remoting-bisocket-service.xml.
Is this ok? It is strange that the messaging is using some configuration inside $JBOSS_HOME/docs/examples.
Anyway, this happens (the warning message) when I stopped the first long-live-consumer (the one that should receive the messages, but it does not), so I suppose that this is normal and that it is not related to the problem.
So then it is not because of the debug mode too..
What can I do more to investigate this further?
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4125591#4125591
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4125591
18 years, 2 months
[JBoss Seam] - Re: What are the aslias defined for the built in seam compon
by norman.richards@jboss.com
I haven't benchmarked this, but the lookups are basically a series of hashtable lookups, which makes it a pretty speedy thing altogether.
On the issue of creating a name index for components, the problem I see is that the underlying contexts are very dynamic. The meaning of "foo" could easily change out from underneath us. (if there weren't the potential for conflict, we wouldn't need namespaces in the first place) I can't see any way that the required indexing (even if it were possible for all contexts) wouldn't be a much more expensive operation.
I think the best solution is that if performance is a concern, always use the fully qualified name or create aliases. In the core seam code we always use the fully qualified name. Failing that, @Import and your own custom factory values could be used to optimize lookups for anything you observe to be a performance problem. So far, of all the inefficient things Seam is doing, this is fairly low down the list of things that are likely to have an impact on overall performance.
That being said, we love to see profiling results on real applications. We are 100% committed to making sure Seam isn't slowing your application down.
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4125587#4125587
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4125587
18 years, 2 months