Re: Load test; v$session.status=ACTIVE and v$session_wait.event=SQL*Net message from client

Okay - that was the easy guess.
The next guess is not really provable.

If you run your query through autotrace, you may find that the plan for joining v$session_wait to v$session involves a sort merge between the two sets of data.

Since the sorts occur in sequence, a little time has to elapse between one set of data being queried and the other being queried. V$ are not read-consistent.

Possibly you get all the sessions which are active, and then start looking for matching session_wait rows - and by the time you start looking, the sessions which were active have gone inactive and are waiting for something.

By blasting a load of fast queries at the database, I managed to get an occasional 'ACTIVE' session that was also 'WAITING'. You might find it easier if you have a lot of busy sessions running lots of very short round-trips.

-- 
Regards

Jonathan Lewis

http://www.jlcomp.demon.co.uk/faq/ind_faq.html
The Co-operative Oracle Users' FAQ

http://www.jlcomp.demon.co.uk/seminar.html
Optimising Oracle Seminar - schedule updated Sept 2nd





"NetComrade" <netcomradeNSPAM_at_bookexchange.net> wrote in message
news:4136577e.621818928_at_localhost...


> Jonathan,

>

> Thank you for pointing this out.

> I've taken a look at the 'STATE' column, and it's 'WAITING' state.

> What do you think this would indicate? It only showed 'WAITED KNOWN

> TIME' for the current session (session running this query).

>

> On Wed, 1 Sep 2004 19:45:34 +0000 (UTC), "Jonathan Lewis"

> <jonathan_at_jlcomp.demon.co.uk> wrote:

>

> >