HELP!! Can a dba control how often PMON purges sniped/killed sessions??

From: Richard A Papaj <papaj_at_acsu.buffalo.edu>
Date: Wed, 1 Sep 1999 11:09:18 -0400
Message-ID: <7qjf44$c5r$1@prometheus.acsu.buffalo.edu>

Can a dba control how often PMON "wakes up" to eliminate sniped or killed sessions? I've seen quite a variance in how often PMON cleans them up. I see them lingering as killed or sniped via v$session.

In a 7.3.3.6 instance, we have an app that's leaving around erroneous sessions for the one userid it uses to connect. It doesn't seem to disconnect from the db properly and the sessions just go inactive after a quick query.

Until it's debugged & resolved, I've increased "sessions per user" in the profile for that userid to avoid reaching the maximum sessions per user which will then lock out that userid from further connections. Also, I've decreased "idle time" in the profile to have Oracle "snipe" the inactive sessions on a more timely basis.

Even though Oracle is sniping the sessions sooner, I'd like to see PMON get around to cleaning them up more quickly. If I'm not mistaken, a sniped or killed session still counts toward the "sessions per user" limit.

I'd appreciate any insight you can provide...

-Rick Papaj Received on Wed Sep 01 1999 - 10:09:18 CDT