On Tue, Apr 1, 2008 at 11:46 PM, Rick Troth <[EMAIL PROTECTED]> wrote:
> > The goal will be to put the identified guest to 'sleep' until we can > > evaluate if it is truly causing issues. ... > > I think I have followed this whole thread, > and I don't seem to recall either of these suggested: > > cp send cp thebadguest sleep Very true. I believe the motivation for the alternatives is based on wishful thinking that when you give the server only a tiny amount of CPU time it will use that to manage network connections and other important things, and not spend it on looping... When you squeeze it hard enough things will be broken anyway when you get at it, just like some aspects will have changed when you wait long enough (e.g. clients have moved on). Rob PS There is also a lot of value in a performance monitor that can tell you when the problem started and will let you review interesting numbers from that time... -- Rob van der Heij Velocity Software GmbH http://velocitysoftware.com/
