On 22 sep 2009, at 20:39, Shawn Drew wrote:

Hehe, just as a thought experiment...

Presuming our instances are of equal size, they would be equally
running
jobs at the same times.  So they would all equally have a couple
drives
available to be pre-empted.

statistically speaking, yes, on average....

In reality, our instances are not equal in size and it is the large
ones
that would tend to have jobs running during the day.  (Smaller
instances
finish by 7am)
But also, since they are large, they are also more likely have the
data
someone would want restored.  Which means they would also have drives
available to be pre-empted.

again statistically.....

And then there are people with opinions on statistics: there are three
kinds of lies... ;-) And of course, there is a guy called Murphy.

On average, you wind up either running into problems, buying too many
drives (which I guess IBM&HP&SUN/Oracle don't mind) or finding a
software solution outside TSM.

I supposed I'm a TSM geek since I think this is fun!

It is fun. Technology can be a beautiful thing....

Regards,
Shawn
________________________________________________
Shawn Drew

--

Met vriendelijke groeten/Kind regards,

Remco Post