> Assume a server supports a set of HTTP driven applications - a > browser or a > soap client of some sort - that are all coming through a sticky load > balancer. I assume a single JNDI tree so that all of the nodes in our > cluster have the same world view so to speak. Is that assumption > correct? A
yes, thanks to ha-jndi. > second question is what happens when you have a set of unrelated > application > server clusters or external applications using the same data database? > Specifially what is the impact on cluster A when records are adding or > removing through cluster B? Is this a pencil-eye issue (mom, it > hurts when I > stick this pencil in my eye:)? well, I suggest you take a look at the clustering documentation. You will have to either use commit option B/C to go to the database each time (and use pessimistic locking at the db level i.e. select for update trick i.e. <row-locking> tag) or use a cache invalidation framework. Distributed cache will come in 4.0. cheers, sacha ------------------------------------------------------- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! http://www.vasoftware.com _______________________________________________ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
