Re: Is an INACTIVE session executing a statement?

From: Martin Berger <martin.a.berger_at_gmail.com>
Date: Thu, 20 Feb 2014 12:41:40 +0100
Message-ID: <CALH8A90YysQVkOWstAuLqjShCFT-iQZayYTHQvnPngSenw9eUw_at_mail.gmail.com>



Dimitre,

I for sure have no transaction (v$transaction is empty) but I can see the sql_id in v$session;
at the same time I see a SQL_EXEC_ID in v$open_cursor. For me it looks like a coincidence of caching or similar effects.

Right now I'm searching for a way to see if there are differences in mutexes the session holds on the cursor - v$latch_children is not populated for cursors in 11.2.?.?.

I will keep you updated ;-)
  Martin

On Wed, Feb 19, 2014 at 10:05 PM, Radoulov, Dimitre <cichomitiko_at_gmail.com>wrote:

> Hi,
> I'm quite sure I've seen an inactive session with not null sql_id in
> v$session.
> I believe those had active transactions (before commit/rollback was
> issued).
>
>
> Regards
> Dimitre
>
> Il 19/02/2014 19.01, Martin Berger ha scritto:
>
> after some testing with Kim Berg Hansen I can only say I'm having a
>> strange observation:
>> in my env v$session keeps (sometimes) SQL_ID even the statement is closed.
>> We compared v$session view for 11.2.0.4 and 12.1 but there is no
>> substantial difference there.
>> Even Kim could not reproduce the problem, he always saw the SQL_ID
>> disappear after the statement completed.
>>
>> Maybe I'm hunting a different problem; at the moment I can not give more
>> details to hunt down my observations.
>> Martin
>>
>>
>

-- 
Martin Berger           martin.a.berger_at_gmail.com
Lederergasse 27/2/14           +43 660 660 83306
1080 Wien               http://berxblog.blogspot.com

--
http://www.freelists.org/webpage/oracle-l
Received on Thu Feb 20 2014 - 12:41:40 CET

Original text of this message