Re: uuid

Tom,
>
> 'Useless' in what way? It should still be a 'unique identifier' which is
> its fundamental purpose. I guess I could imagine a small troublesome time
> window where two systems could be rapidly blasting out uuids concurrently
> during the time where adaptor addresses changed in such a way that the
> address of one briefly duplicated the address of the other and the two
> clocks were far enough out of sync; but a change of adaptor address implies
> possibly ending and restarting many services anyway and that can imply
> ending and restarting numerous applications.
>
> If addresses don't overlap and clocks are reasonably maintained, there
> shouldn't be any problem.

I dont mean that the adaptor addresses are changed on the fly, but that they
exist at customer sites and I have no control (or knowledge/interest) in
them... ie in all the AS400s in the world, how many have the same address
because of the old reverse-addresses?
I think whether uuid uses the internal address of the card (which it clearly
should) or whether it uses the line description is so fundamental to the
point of uuid that the documentation should be clear on the matter. As it is
the only definitive answer would require inside knowledge...