Re: Cancelling User sessions

From: WolfAlpha <wolfalpha_spamguard_at_home.com>
Date: Mon, 05 Jul 1999 05:59:19 GMT
Message-ID: <XQXf3.12558$5i7.7598@news.rdc1.va.home.com>

Please excuse this reply if a similar answer has already been posted (I only caught the tail end of the thread).

You could have your DBA modify the default profile (or create a new profile and assign it to all users) that sets the IDLE_TIME resource to some value other than unlimited. This resource restriction will cause any user connection to be disconnected after being inactive on the system for any amount of time greater than this value. Hope this helps.

Jeff S

Bhooshan Prabhu <bhooshan.prabhu_at_citicorp.com> wrote in message news:377C938C.23D3FDF6_at_citicorp.com...
> Hi Nigel,
>
> A session is killed by issuing command "ALTER SYSTEM KILL SESSION 'x,n'".
> For a user with non-dba privileges to achive this, each such user will
have
> to have a ALTER SYSTEM privilege granted to her / him. I don't think, your
> DBA will agree to that.
>
> Your DBA can write a piece of code, which can look for sessions, that have
> been inactive for a acceptable time limit i.e. 30 minutes or so and issue
> statements to kill these sessions. This piece of code, can then be
submitted
> as a dbms_job so that Oracle will execute it, at intervals as specified
> during job submission.
>
> Regards
> Bhooshan
>
> --------------------------------------------------------------------------

---

>
>
>
> Nigel bruno wrote:
>
> > We have a problem at the moment were sessions are left hanging and
> > causing other sessions to lock out. Because we don't have the privileges
> > to kill sessions we currently have to contact the DBAs to do this for
> > us, which obviously takes too much time.
> >
> > Is it possible to create a procedure that will allow a user to cancel
> > sessions with the same user id as the user executing the command. This
> > will allow us to cancel our own sessions without causing a security
> > breach.
> >
> > I think I know how to code it but could someone confirm the theory is
> > OK. I'm assuming as long as the owner of the procedure has kill session
> > privilege then users who have access to the procedure will be able to
> > kill the sessions belonging to them. The procedure will check that the
> > user id of the session to be canceled is the same as the user id of the
> > user running the procedure before the session is killed.
> >
> > Hope this make sense.
> >
> > PS Any ideas why debugging VB and stopping the program causes sessions
> > to be left open.
>
>
>
Received on Mon Jul 05 1999 - 00:59:19 CDT