Well, if you have parallel_automatic_tuning set to true, your server processes will utilize all the server processes it has available to them up the the upper bound of parallel_max_servers. If there are multiple server processes requesting parallel processes then the number they request may be different to the number that they receive, its as simple as that. Post the output of "show parameter parallel" here.
Now check the number of parallel processes being used at any one time (select sid, req_degree "requested_dop", degree "actual_dop" from v$px_session order by sid; ). You might find that you don't have the maximum set high enough for your demand.
Last edited by waitecj; 11-08-2006 at 06:32 AM.
Reason: Stoopid emoticon at the end of "sid"