Re: Strategy for "one-off" tasks

[Daniel Pittman]

On Fri, 11 Jul 2003, Christopher DeMarco wrote:
> On Thu, Jul 10, 2003 at 03:34:06PM -0700, Darren Dunham wrote:
> 
>> I think the idea is "why  would you need  to perform a one-time-only
>> task"?  Can you give an example of what this would be?  Why would it
> 
> I'm using cfengine not as a means to converge many diverse systems to
> one Ideal Functional State, but to painlessly install and update a
> bunch of identical, dumb, kiosk-type units. So killing X is a nice
> single-function action that I'd like to perform remotely; passwordless
> ssh would work except that I already HAVE cfengine.

I think that the idea that killing X is a one-time task is wrong.

If your policy is that a manual decision is needed to kill X, in which
case a hand-defined class and cfrun will do it.

If not, X should never be running. Having cfengine check every time it
executes that X is not running, and killing it if it is, is not a time
consuming operation and ensures that you *never* have X running...


The time taken to check something is right is minimal if the computer
does it automatically.

The risk of assuming that things never change is much higher, though,
since that actually takes *your* time.

      Daniel

-- 
A wonderful discovery, psychoanalysis.  
Makes quite simple people feel they're complex.
        -- S. N. Behrman