Re: determining bind values in deadlock situations

  • From: "Joel Wittenmyer" <joel.wittenmyer@xxxxxxxxxxxxxxxxx>
  • To: "oracle-l@xxxxxxxxxxxxx" <oracle-l@xxxxxxxxxxxxx>
  • Date: Thu, 05 Mar 2009 13:58:23 -0600

'Application Error' is just one of many reasons you can get a deadlock.
 The bind variables help you to understand exactly what the application
was doing.  Even if it is in fact an application errors, the bind
variable values in conjunction with the rest of the information in the
Process State often helps the developers figure out exactly how they got
themselves into that situation.  It can also tell you if the application
caught the error and re-tried the statement (if a DML) or not.  Which is
to say that you can tell if you have lost data.  By laying out the
entire sequence of events that one process did, along with the bind
variable values, for the VP in one company we resolved a deadlock that
the company had been plagued by for years and had given up hope of
resolving.  For another company, I showed them that they were losing
about 100 records / day.

> 
> On Fri, Feb 20, 2009 at 5:33 PM, Barun, Vlado <Vlado.Barun@xxxxxxx> wrote:
> 
> >  Environment: 10.2.0.3 RAC/ASM on Solaris 10
> >
> > I?m trying to find an efficient way to get the bind values for
statements
> > that are involved in a deadlock situation.
> >
> >
> Maybe I am just being particularly dense today, but I can't think of why
> you would need bind variables to resolve a deadlock.
> 
> The ORA-60 deadlock is caused by an application error, typically two
> different parts of an application causing a deadly embrace by trying
> to lock the same tables, but in different orders.
> 
> The trace file generated by the ORA-60 contains the SQL and the
> rowid's of the affected rows.
> 
> Is there something I am missing?
> 
> Jared Still
> Certifiable Oracle DBA and Part Time Perl Evangelist
> 
> 

-- 

--
//www.freelists.org/webpage/oracle-l

Other related posts:

  1. Home
  2. oracle-l
  3. Archive
  4. 03-2009