In today's Forums, rickcarson follows up on an open question from yesterday and makes the case for Jini in Re: Java Cluster, whats the best way? "Jini is a really good way to do a clustered supercomputer. What you do is on each of the computers you put a Jini client. Then you expose a Jini service. What happens is that the client goes to the service and asks for work. The service gives it some work to do. The client does the work. The client goes back to the service to ask for more work. Jini is nice in this case because it has the concept of expiry. Eg if one of the clients bursts into flame its lease will expire, and the service can figure out that if the lease expires whatever piece of work it last handed to that client needs to be handed to a different client. Moreover, real users can be using the computers for other stuff, and you can run the Jini clients highly niced and in the background, just using the spare cpu cycles without stealing from the users."

Joe Shevland offers a reality check on JPA annotations in RE: EJBQL Exception. "I'd say though the cases where you don't need to define a property->column mapping are pretty rare (almost restricted to tutorials and shallow examples I'd say) - all projects I've worked on to date have either had a legacy database, or a schema that's being dictated by DBA's/database standards for column names etc. A JPA-generated database schema is probably not going to be very good if you don't use the Column mapping for things like length, nullability etc. Or on the other hand, you'd have getters/setters like getPrs_Fst_Nm() which I personally wouldn't allow in a codebase."

hunterp reports a seemingly pathological garbage collection problem in GC killing me. "I have an app and GC is killing me. everything is fine until the memory reaches the heap then gc takes forever (while still applying load) and the app is completely unresponsive (up to 2 minute wait times). I thought the time ratio combined with incremental would collect more aggressively (i have lots of free cpu). But with the parameters below, for a 4 minute run, the jvm spent well under a minute GC before it hit the wall and blew up. FYI the PSOldGen space is where 75% of the garbage stays until the wall. I've tried the -XX:MinHeapFreeRatio=30 (instead of default 40) but no difference."