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: RMAN, v$session_longops and Ora-03113

Re: RMAN, v$session_longops and Ora-03113

From: Tim Gorman <tim_at_evdbt.com>
Date: Wed, 18 May 2005 11:07:42 -0600
Message-ID: <BEB0D27E.27801%tim@evdbt.com>


RMAN posts to V$SESSION_LONGOPS extensively, presumably (but not positively) using the DBMS_APPLICATION_INFO packaged procedures as anybody can...

That's the only way the two can be linked. Still unsure about how posting to V$SESSION_LONGOPS by RMAN can cause ORA-03113 on subsequent queries by other sessions -- as Jared indicates, it sounds rather tenuous to associate one to the other. It smacks of the logic that runs "Whatever changed recently must be the cause ... somehow or other"...

on 5/18/05 10:35 AM, Jared Still at jkstill_at_gmail.com wrote:

> Hi Tim,
> I am confused about the connection between v$session_longops and
> making a backup. 
> 
> Your email asserts that the v$session_longops problem prevents
> backups, but I don't see the connection.
> 
> Jared
> 
> 
> 
> On 5/16/05, Tim Onions <att755_at_hotmail.com> wrote:

>>
>> 8.1.7.4.18 under Windows 2000 sp4 using Veritas Netbackup (Apologies if
>> this
>> did get posted already but I received a failure notification so I assumed
>> it
>> had not.)
>>
>> Dear All
>>
>> We are seeing far more regular issues with v$session_longops causing a
>> ORA-03113 when you try and select from it since we upgraded to the
>> terminal
>> release of 8.1.7 and at roughly the same time moved from Tivoli to Veritas
>> for our RMAN backup back-end.
>>
>> Metalink says fixed in 9.2 which we will go to in due course but not for a
>> few more months (I know we should be there already but....). Other than
>> that
>> it says you have to bounce your DB which we can only do at weekends - so
>> if
>> the problem occurs on a Monday we are without backups all week.
>>
>> I would like to know if anybody has any knowledge of what causes
>> v$session_longops to go "bad" (so I can avoid it), or how to work around
>> it
>> when it is bad (Metalink says try setting hash_joins_enable FALSE but that
>> does not work for me). I see one backport request for a fix in Metalink
>> but
>> nothing to say it ever got done or if so for my platform (probably not as
>> I
>> rasied a TAR and they denied any existance of a backport) - so any
>> knowledge
>> of a backport would be appreciated too.
>>
>> Many thanks
>>
>> Tim Onions - Oracle DBA
>>
>> _________________________________________________________________
>> Winks & nudges are here - download MSN Messenger 7.0 today!
>> http://messenger.msn.co.uk
>>
>> --
>> http://www.freelists.org/webpage/oracle-l
>>

>
>
--
http://www.freelists.org/webpage/oracle-l
Received on Wed May 18 2005 - 13:12:42 CDT

Original text of this message

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