RE: determining bind values in deadlock situations

  • From: "Barun, Vlado" <Vlado.Barun@xxxxxxx>
  • To: Tanel Poder <tanel@xxxxxxxxxx>, "'Bobak, Mark'" <Mark.Bobak@xxxxxxxxxxxx>, "oracle-l@xxxxxxxxxxxxx" <oracle-l@xxxxxxxxxxxxx>
  • Date: Wed, 4 Mar 2009 07:57:07 -0500

My statement was not correct :) 

Thank you for educating me on this.

Would hanganalyze_global level 4 provide the all statements involved in the 
deadlock? 
If not, is there a way to get all the statements?

Regards,
 
Vlado Barun, M.Sc.
Sr. Manager, Database Engineering and Operations
Jewelry Television
Mobile: 865 335 7652
Email: vlado.barun@xxxxxxx


-----Original Message-----
From: Tanel Poder [mailto:tanel@xxxxxxxxxx] 
Sent: Wednesday, March 04, 2009 7:52 AM
To: Barun, Vlado; 'Bobak, Mark'; oracle-l@xxxxxxxxxxxxx
Subject: RE: determining bind values in deadlock situations

Vlado,

That is not correct ;)

What you see in lmd trace below is two copies of the same SQL for one
session and two copies of the same SQL for the other session. 
As the trace data itself says it's the "Current SQL" just dumped twice, not
"past SQL" or "SQL doing the locking".

Run your test case with 4 *distinct* SQL statements and you'll see only 2 of
them, the currently executed ones are shown.

--
Regards,
Tanel Poder
http://blog.tanelpoder.com



> -----Original Message-----
> From: Barun, Vlado [mailto:Vlado.Barun@xxxxxxx] 
> Sent: 04 March 2009 06:28
> To: Bobak, Mark; tanel@xxxxxxxxxx; oracle-l@xxxxxxxxxxxxx
> Subject: RE: determining bind values in deadlock situations
> 
> That is not correct.
> 
> Sql_a, sql_b, sql_c and sql_d are in the lmd0 trace files in 
> a RAC environment, assuming you have the necessary patches 
> applied (in my case 4441119 for 10.2.0.3 RAC/ASM on Solaris 10).
> 
> 
> Following is a test case:
> 

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


Other related posts: