On Wed, Jul 29, 2009 at 10:49 AM, Ted MacNEIL <[email protected]> wrote:
> >But also a cpu use limit per class is another > way to catch a looper early. > > But, what if it's not looping? > Exactly. Most programs actually DO loop, but eventually they finish processing whatever it is they're processing and they end. So the hard part is knowing whether an apparently "looping" program is in trouble or just doing what it was intended to do. There is also the question of potential down-stream effects of cancelling one of these. Are there any dependencies? Can you just nuke it and hope for the best? With test jobs, arguably "yes". With production jobs... erm... well usually I suppose, but often enough "not". If it was my ass on the line, I would want the application owner to either make the decision, or have a policy of some sort in place to tell the operator (or automation) what action to take. -- This email might be from the artist formerly known as CC (or not) You be the judge. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
