On Tue, 2008-09-23 at 22:17 -0700, Joshua D. Drake wrote:
> Simon Riggs wrote:
> > On Tue, 2008-09-23 at 16:50 -0400, Andrew Dunstan wrote:
> >
> >> If we get all that done by November we'll have done well. And we know
> >> that in some cases just this much can lead to reductions in restore
> >> time
> >> of the order of 80%.
> >
> > Agreed. Go for it.
> Just as an FYI, by far the number one bottle neck on the multiple work
> restores I was doing was CPU. RAM and IO were never the problem.
It would be useful to see a full breakdown of those results.
There's always a bottleneck on something for any particular task and we
shouldn't presume the problem is only on CPU, for all data on all
systems. CPU parallelism is the most pressing problem, I agree, but I
think we will quickly hit problems without memory limits. But I agree
with Andrew that this will be a nice problem to have and not everything
is possible by Nov 1.
-- Simon Riggs www.2ndQuadrant.comPostgreSQL Training, Services and Support
