View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3820730#3820730
Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3820730 Yes, you can help. The current HA-JMS implementation leaves it up to the client to register a Connection Exception listener. When the current master singleton server fails, the client will receive a callback with this exception, at which point it has to reconnect again via the HAConnectionFactory in order to get the new master singleton server. You can try to implement the logic in the client portion of the HAIL transport that keeps track of the changes in the back end and automatically fails over to the new master server. There are at least two ways to do this. 1) The preferred one is to reuse the latest work done in the JBoss detached remote invokation layer. This way you will effectively acomplish two pending tasks, one is improving the client fail-over, the other is unifying the invokation layers. 2) Another way to do the client fail-over is to borrow logic from the smart RMI proxies, which keeps track of the available servers. In either case you will need to add extra logic to the client to stay informed about the latest master server. Not a piece of cake. Quite a bit of work and research ... Let me know if you will be working on this. Ivelin PS: A caveat is that this works for non-durable sessions. If a session is durable there needs to be even more work done to ensure that there is no message loss during fail-over. This a problem that can be taken care of separately on the server side though. ------------------------------------------------------- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn _______________________________________________ JBoss-Development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development
