On Thu, 2002-11-07 at 19:36, Christopher Kings-Lynne wrote:
> Even nastier, what about if the different postgres servers in the cluster
> run on different architectures! That way you'd get different floating point
> results on each machine...
Oh come on! There's no such thing as "perfect" anywhere. I was the one
that originally posted the concept that one could maintain identical
databases by piping queries to two DBs simultaneously. It's clearly
been demonstrated that this is not possible in the general case.
Differing architectures for the servers is the *least* of the problems.
As always, "don't insure what you can afford to lose" and the converse,
"always insure what you can't afford to lose" both apply.
b.g.