Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> Re: APPNAME in trace

Re: APPNAME in trace

From: Edgar Chupit <chupit_at_gmail.com>
Date: Thu, 8 Jun 2006 07:55:34 +0300
Message-ID: <a8f0771c0606072155h5192a7c8n891ec63793dcb190@mail.gmail.com>


Nelson,

Please note, that in 10g you don't have to call dbms_session.set_module to "register" you module in the database (tracefile). 10g allows you to piggyback module information using "End to End Metrics", for example, in java (this applies to all Oracle Call Interfaces) you can use OracleConnection.setEndToEndMetrics() and Oracle will not send message instantly, but will piggyback this information to next server call (parse/execute/fetch/etc).

Using End to End Metrics you will remove one roundtrip between your application and database server, but in your case, you can remove many roundtrips using this technique.

You probably are aware of this, but I still decided to make a note.

On 6/7/06, Petersen, Nelson <nelson.petersen_at_homehardware.ca> wrote:
> Thanks James and John.
>
> The PL/SQL, developed in-house (not Oracle Apps), is called from a batch job
> which runs a script.
>
> The script doesn't loop, but the PL/SQL does ... through all
> order_lines for a specified order_release_id.
>
> I just didn't understand why consecutive "APPNAME mod=" lines would be
> emitted in the
> trace. I've had a heart-to-heart with one of our support analysts.
>
> Standards (established for good reasons) led to the use of SET_MODULE.
> For this particular batch job, *every* procedure and function call registers
> and de-registers
> the module or action. Functions are called during assignment operations.
> The trace file doesn't show assignments (may level-12 would?), so it was not
> really obvious what
> was calling SET_MODULE.
>
> I'm having a little trouble with calculating how much this contributes to
> response time,
> but I'll review Cary/Jeff's book tonight. Some time seems to be missing,
> but I hope
> to pinpoint that tonight.
>
> Thanks again,
> Nelson
>
>
> -----Original Message-----
> From: John Kanagaraj [mailto:john.kanagaraj_at_hds.com]
> Sent: Wednesday, June 07, 2006 15:05
> To: James.Foronda_at_Sun.COM; petersen_n_at_srs015.homehardware.ca
> Cc: oracle-l_at_freelists.org
> Subject: RE: APPNAME in trace
>
>
> Nelson,
>
> Is this an Oracle Apps (E-Business Suite) instance? In that case, is
> this from a Concurrent report? You may not be able to see the SET_MODULE
> as the wrappers around the Concurrent report may have set this....
>
> John
>
> -----Original Message-----
> From: oracle-l-bounce_at_freelists.org
> [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of James Foronda
> Sent: Wednesday, June 07, 2006 8:44 AM
> To: nelson.petersen_at_homehardware.ca
> Cc: oracle-l_at_freelists.org
> Subject: Re: APPNAME in trace
>
> Nelson,
>
> Petersen, Nelson wrote:
> > Would the trace be writing these lines out-of-order?
> > I've looked at the PL/SQL (developed in-house) and I don't see
> > consecutive calls to SET_MODULE.
> >
>
> Is there only one call to SET_MODULE in your PL/SQL code? If yes, is
> that code being called from outside of Oracle (some scripts or GUI
> program)? Maybe the calling program loops around the call to the PL/SQL
> code?
>
> James
> http://jforonda.blogspot.com
> --
> http://www.freelists.org/webpage/oracle-l
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

-- 
Best regards,
  Edgar Chupit
  callto://edgar.chupit
--
http://www.freelists.org/webpage/oracle-l
Received on Wed Jun 07 2006 - 23:55:34 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US