- I've added a slight improvement to the multicast detector for caching the detection notification and serialization bytes and only re-serialization if the detection data is different. - the notification cache has an improvement where it will only make one connection to a remote server for async notifications in the event multiple listeners are added.
- I added a meaninful name to each thread pool worker thread in the JMX ThreadPool utility class to make it easy during a thread dump to tell what the thread is
- I fixed a couple of problems found when doing failure testing of multiple server/clients.
- Replaced the NetworkRegistry notifications to use a thread pool vs. creating new threads
- I added a listener for detection of remote server failure and destroy the local client proxy to the remote mbean server automatically upon failure.
These fixes have been committed to the 3.2 branch. Talking with Tom, he suggested (which I agree) that we wait until we have an official next release before committing these changes to HEAD.
I plan to continue running JProbe against both remoting and jmx-remoting in the next week as we continue to integrate the 3.2.4 release into our product and hope to work with Tom to add some more scalability improvements soon. I'm particular interesting in continue to evolve/improve the async transport since I think it has a lot of performance optimizations in cases where a response isn't needed.
Jeff
------------------------------------------------------- This SF.net email is sponsored by: IBM Linux Tutorials. Become an expert in LINUX or just sharpen your skills. Sign up for IBM's Free Linux Tutorials. Learn everything from the bash shell to sys admin. Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click _______________________________________________ JBoss-Development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development
