OraFAQ Forum: Server Administration » last_call

Home » RDBMS Server » Server Administration » last_call_et column on v$session

Show: Today's Messages :: Polls :: Message Navigator
E-mail to friend

last_call_et column on v$session [message #59222]

Wed, 05 November 2003 03:16

Celia
Messages: 8
Registered: April 2003

Junior Member

Hi I am trying to find the definitive length of time that a session has been idle.
What measurement is last_call_et in? Is it seconds, minutes etc and how often is this field updated for INACTIVE sessions?
On all the FAQ I've read, it seems to suggest that last_call_et is in seconds but 15 minutes after running some sql, the last_call_et value for my session still says 3.

Report message to a moderator

Re: last_call_et column on v$session [message #59223 is a reply to message #59222]

Wed, 05 November 2003 04:16

Maaher
Messages: 7065
Registered: December 2001

Senior Member

Tom Kyte to the rescue...again. Amazing what this guy has done. Anyway, here you go.

Put asktom.oracle.com in your favourites. It's worth it.

MHE

Report message to a moderator

Re: last_call_et column on v$session [message #59225 is a reply to message #59223]

Wed, 05 November 2003 04:55

Celia
Messages: 8
Registered: April 2003

Junior Member

Thanks. Yes I have similar code but my question was really whether last_call_et was always guaranteed to be the number of SECONDS of idle time and how often Oracle updated this field. I had a test idle session for over 20 mins and last_call_et still said 3.

Report message to a moderator

Re: last_call_et column on v$session [message #59229 is a reply to message #59225]

Wed, 05 November 2003 05:59

Maaher
Messages: 7065
Registered: December 2001

Senior Member

No, I don't think it is guaranteed to be updated frequently, like Tom said one of his pages (quote):
----------------------------------------------------------------------
Note: the last_call_et column might not be updated frequently if you are not
using timed statistics. It'll get updated about every 40 seconds or so.
----------------------------------------------------------------------
(end quote)

But I've just seen a Metalink document referring to a Bug (#1540012) which might cause this incorrect value. The bug was confirmed for 8.1.7.4 and was believed to be reproducable for earlier versions (every OS) too. Unfortunately it's an internal bug, so not accessible to the public. It should be fixed in 9i.

MHE

Report message to a moderator

Previous Topic:	scheduling/Thiru
Next Topic:	table partitioning

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

]

Current Time: Fri Sep 20 08:34:05 CDT 2024