Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Mailing Lists -> Oracle-L -> Re: ORA-600 Deadlock Issues
On 06/24/2005 09:59:21 PM, Karthik wrote:
> Hello,
>
> I have code that consistently bombs due to ORA-600 - Deadlock Detected.
> The code base is very complex, and a bunch of tables get updated via
> direct updates/triggers etc., so debugging at the app level is proving
> nightmarish.
>
> Is there any way I can get a hint (from alert.log or elsewhere) where
> the contention is?
>
> Thanks for your help.
> Karthik
You have a problem with the CPU, which is multiplying the error code
by 10. Let me guess: Intel inside?
Deadlock used to be ORA-0060. ORA-0600 is a little bit, hmmm, different.
It ain't better, that's for sure.
How to handle deadlocks? Well, one way of seeing what the problem is is
to do the following:
SQL> alter session set events='60 trace name errorstack forever, level 12';
Session altered.
SQL> If and when you encounter error IN THAT SESSION (you may well encounter it in another session), you will get a trace specifying everything. If you'd like it to apply to the whole system, you may try "alter system set events" instead or, if you have a SYSDBA connection, the this:
SQL> oradebug event 60 trace name errorstack forever, level 12
Statement processed.
SQL>
Also, you can try with hanganalyze. It looks like this:
SQL> oradebug hanganalyze 3
Hang Analysis in /oracle/product/10g/admin/oracle/udump/10g_ora_7521.trc
SQL>
You can find more about the hanganalyze utility in the note 215858.1 or,
if I'm not mistaken, in NYOUG publication from June.
-- Mladen Gogala Oracle DBA -- http://www.freelists.org/webpage/oracle-lReceived on Fri Jun 24 2005 - 22:36:25 CDT
![]() |
![]() |