posted 18 years ago
Hi Sitaram,
Admitting that we don�t play other container specific trickeries (like allow-concurrent-calls SFSB specific tag with weblogic) I could see two options:
The container blocks the current transaction, and try to passivate (using either LRU or NRU) in order to release the cache. If the cache could not be released before transaction times out, an exception will be thrown.Throws kind of cache full exception right away.
Can you think of other options?
First doesn�t make much sense, because the blocked transaction could hold valuable resources. If this happens for many simultaneous transactions it could have pretty bad consequences. Hence I�ll go with the second option. Of course this is just my opinion (which I never tested, since in practice I never used SFSBs that often).
Regards.
I think, therefore I exist -- Rene Descartes