posted 20 years ago
The client can call a method (on a CMT bean), and it might start a transaction. The bean might even call other methods that start other transactions (but no nested transactions, of course). In the end, when that original method from the CMT bean ends, there must be no transactions open. The container manages that. Think about your DD.... you had to tell it to use CMT for that bean. You also had to tell it how you wanted it to handle transactions (Required, Not Supported, etc). There's nothing in the DD to say "leave this one open at the end of the method", right?
--Dale--