TRacing in shared server

Question

Hi when i enable session level tracing in dedicated process the trace is created in formated ora_sid_spid..........
But how is it gennerated in the case of shared server architecture..Is there any difference..If so how to find trace files

Answer 1

trcsess is to read the multiple generated trace files after you have enabled tracing for a session and then create a combined trace file as the output file.
Did you enable tracing ? Using "normal" methods to enable session tracing ?
Then, did you see the trace files generated in the dump_dest ?
Then, you trcess.
(Note : If your user_dump_dest has many trace files in the same directory, trcsess will take some time going through all of them to match the "SID.SERIAL#" combination. Also, if there are some older trace files which just happen to have the same "SID.SERIAL#" combination they will, wrongly, get included because trcsess matches only "SID.SERIAL#". {It would wrong to say "I expect it to match the timestamps , becasuse it doesn't know when you started tracing and stopped tracing : If it finds a trace file dated 01-Jan and trace files dated 10-Jan and 11-Jan, all for the same SID.SERIAL# combination, it is logical for it to assume that  they all belong to the same session, traced over a period of 11 days !}
Therefore, it is also sensible to identify the trace files in advance and provide a list of trace files on the command-line so that trcsess ignores the other trace files)
Also, since 10g, you can use DBMS_SESSION.SET_IDENTIFER in your application code and have trcess filter by the CLIENT_ID rather than by SID.SERIAL#
Hemant K Chitale
Edited by: Hemant K Chitale on Jan 12, 2011 2:56 PM
Edited by: Hemant K Chitale on Jan 12, 2011 3:05 PM