Eric
Oh, that sounds good. How does the implementation know/decide whether it
has to parse the whole message or just the header?
Basically the parser is only called as the tree is walked. So if no
logic needs to walk the tree, then it will only be streamed through. I
believe that the XPath code asks for the first matching element, so it
should stop even without the [0], but its worth being careful.
a cluster - basically you "gracefully shutdown" each server in turn,
with the result that no requests are lost, and the load balancer
redirects work to the rest of the cluster.
"Gracefully shutdown"? I hope you mean "gracefully restart" each
server...? ;-)
The Graceful shutdown stops accepting new work while the old requests
are processed. The result is that the load-balancer will spot the server
is down and start directing new work to the other cluster members. Once
the server has finished all running work, it can be restarted with the
updated configuration, and then the load balancer will notice it is up
and start sending new work in.
Is this also part of 1.6
Yes.
, or only the cluster support for caching and
throttling. Are there any infos on the implementation? Do you use
something like jgroups for intercluster communication?
We use Apache Tribes which is what Tomcat uses, but the interface is
pluggable.
That sounds to be a good idea. So you at least would have a default
implementation without external dependencies.
+1
Paul
_______________________________________________
Esb-java-user mailing list
[email protected]
http://wso2.org/cgi-bin/mailman/listinfo/esb-java-user
