All, Please stop guessing. Think about it -- ORA-03135 and ORA-03136 are errors returned from the RDBMS engine, not the SQL*Net listener component. If they were returned by the listener, they would be "TNS-xxxxx" messages, not "ORA-nnnnn" messages. The root-cause is in the RDBMS, so look for root cause within that layer. What I've found really useful is to query the active session history (i.e. either DBA_HIST_ACTIVE_SESS_HISTORY in the AWR if the errors occurred some time ago or V$ACTIVE_SESSION_HISTORY if they occurred recently) and look at what the database sessions are waiting upon. Take the timestamps from the error messages, then just browse the appropriate ASH view for that time period, including about 10-30 seconds prior to the error messages, paying particular attention to columns like SAMPLE_TIME, EVENT, P1, P2, P3, PROGRAM, MACHINE, etc. Usually I find a bewildering variety of wait events occurring, indicating a wide variety of causes, anywhere from "enq: SQ - contention" (i.e. contention on AUDSES$ sequence) to "library cache lock" (i.e. shared pool contention due to under-allocated shared pool), to "cursor: pin S wait on X" (i.e. weird application programming logic at connect-time), to "SGA: allocation forcing component growth" (i.e. resizing SGA memory components). The sources of these errors are so many and varied that you really need to understand the root cause. It could be a tweak of Oracle initialization parameters or it could be correction of toxic behavior by the application. It is better to know first before expending time, effort, and credibility doing anything... and that's where the active session history comes in... Hope this helps... Tim Gorman consultant -> Evergreen Database Technologies, Inc. postal => P.O. Box 630791, Highlands Ranch CO 80163-0791 website => http://www.EvDBT.com/ email => Tim@xxxxxxxxx mobile => +1-303-885-4526 fax => +1-303-484-3608 Lost Data? => http://www.ora600.be/ for info about DUDE... On 9/17/2010 6:18 PM, Sanjay Mishra wrote: -- //www.freelists.org/webpage/oracle-l |