Jan,
I think we should suspend further discussion for now...in summary:
ARC Buffer management is an important new performance feature for 7.5;
the implementation is a good one and should have positive benefit for
everybody's workload. ARC will adapt to a variety of situations and has
been designed to allow Vacuum to avoid interfering with user
applications.
That's the important bit: The implementation notes are detailed; I've
read them a few times to ensure I've got it straight. I am confident
that the situation I described CAN exist with regard to multiple
concurrent queries performing full table scans upon a single large
table. Further debate on that point is continuing because of my poor
explanation of that situation; forgive me. Thanks very much for your
further explanations and examples.
I will take a more practical tack on this now: providing evidence of a
real query mix that exhibits the described properties and quantifying
the effects and their frequency. If it IS worth it, and I accept that it
may not be, I'll have a hack at the very specialised improvement I was
suggesting, for very specific workload types.
Best Regards
Simon Riggs
---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?
http://archives.postgresql.org
