Re: tracking down hidden SQL???
- From: Chris Taylor <christopherdtaylor1994@xxxxxxxxx>
- To: Adric Norris <landstander668@xxxxxxxxx>
- Date: Thu, 17 Oct 2013 09:48:45 -0500
Hey, the other thing you could do, if you know some identifying information
for these sessions, like OSUSER, machine/terminal, client_info etc, you
could setup a LOGIN TRIGGER that enables a 10046 trace and capture any/all
SQL executed in the session.
Some apps (like PeopleSoft) will populate client_info so it makes it easier
to trace one session instead of all the connected sessions. Maybe you have
something available to you that will enable you to create a limited trace
on a few sessions.
HTH
Chris
On Wed, Oct 16, 2013 at 3:33 PM, Adric Norris <landstander668@xxxxxxxxx>wrote:
> That wouldn't surprise me at all, actually. Rather annoying that it's not
> visible, however.
>
> On Wed, Oct 16, 2013 at 2:42 PM, Jonathan Lewis <
> jonathan@xxxxxxxxxxxxxxxxxx
> > wrote:
>
> >
> > Just throwing in a wild guess here - Dan't suggestion is the smart thing
> > to do - but if the prev_sql_id is always for "select silly_thing from
> dual"
> > which you're sure is just a connection pool test then it's possible that
> > the invisible statement is just a "rollback" (or "rollback;" or
> "Rollback"
> > or ... ).
> >
> > Regards
> > Jonathan Lewis
> >
> >
> > ________________________________________
> > From: oracle-l-bounce@xxxxxxxxxxxxx [oracle-l-bounce@xxxxxxxxxxxxx] on
> > behalf of Adric Norris [landstander668@xxxxxxxxx]
> > Sent: 16 October 2013 20:15
> > To: Chris Taylor
> > Cc: oracle-l
> > Subject: Re: tracking down hidden SQL???
> >
> > On Wed, Oct 16, 2013 at 2:05 PM, Chris Taylor <
> > christopherdtaylor1994@xxxxxxxxx> wrote:
> > > Did you try querying [g]v$session PREV_SQL_ID as well for those
> sessions
> > > that are showing the SQL_ID you're interested in? (Kerry mentioned it
> in
> > > that blog post). The prev_sql_id might give you a clue about why this
> > > sql_id is showing up as the current one?
> > >
> >
> > The previous PREV_SQL_ID is always identical, but is unfortunately just a
> > connection pool session-test query.
> >
> > select 2+2 from dual
> >
> > I appreciate the suggestion!
> >
> >
> > --
> > //www.freelists.org/webpage/oracle-l--
> > //www.freelists.org/webpage/oracle-l
> >
> >
> >
>
>
> --
> "I'm too sexy for my code." -Awk Sed Fred
>
>
> --
> //www.freelists.org/webpage/oracle-l
>
>
>
--
//www.freelists.org/webpage/oracle-l
Other related posts:
