Thanks for your quick answer. What I expected using cache coordination was to get the whole cluster invalidate objects in the cache upon a successful write operation. Does the "full" Oracle Toplink provides this feature?
I understand that using optimistic locking I can assure safe and predictable behaviour on writes, but how do I ensure reading stale objects from the cache when running on a cluster with cache enabled? Are there any patterns or suggestions for this with Toplink Essentials?
[Message sent by forum member 'kjozsa' (kjozsa)]
http://forums.java.net/jive/thread.jspa?messageID=236790