Re: Session Kill

On Oct 6, 11:52 am, "lyx..._at_gmail.com" <lyx..._at_gmail.com> wrote:
> >What is the difference of being out of gas because
> >you drove 400 miles or because someone emptied
> >your tank?
> >Michel Cadot
>
> This is why I asked this question as I could not find any diff.
>
> See this linkhttp://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:2...
> in asktom
> as he said something which I could not understand.
>
> Then why people discussed about the scripts for killing the idle the
> sessions when this profile option has been
> availabe. You would find this thing in almost each and every oracle
> discussion group about the killing session by
> script or manually. I have come accross lot of sites where these type
> of practice going on .. anyway ..
>
> Mark - Thanks for your reply.
>
> Regards,
> Jagjeet Singh

Because the sessions don't go away!

Imagine you have a small machine that can handle 400 sessions. 300 people leave on Thursday without logging out, say the network went down. You now have a crapload of phone calls coming at dark o'clock AM Friday as person 101 to 300 can't log on because SMON was too busy rolling back some failed batches and never got around to the killed sessions. That's why some people (at least oldtimers on unix) really believe in killing sessions with the OS and letting PMON sort it out. The unreleased locks and ORA-1555's are other problems, depending on version and exact situation. Older versions were worse.

jg

--
@home.com is bogus.
http://catless.ncl.ac.uk/Risks/24.84.html#subj1