On 24 March 2013 13:38, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Simon Riggs <simon@2ndQuadrant.com> writes:
>> Can we do this now?
>
> No.
>
> We are trying to get 9.3 out the door, not undertake design, coding,
> and testing of entirely new behaviors which will have complex
> performance tradeoffs. This sounds like a great project for early in
> the 9.4 development cycle, but it's way too late to try to do it now.
> (At least if you'd like to ship 9.3 this year. If we're going to
> re-open 9.3 for this kind of development, I have some planner fixes
> I need to make ...)
Yes, there are tradeoffs and people would like to control them rather
than being stuck with one and only one behavour that we know has major
downsides in certain use cases.
I'm proposing a parameter to control this, with the existing behaviour
being still obtainable. This is not a new automatic behaviour. If you
have a new optimiser behaviour that can be controlled by parameter and
takes a few hours to implement, please do it.
I'm greatly concerned that nobody seems to care about replication
delay; this is the second time I've raised major issues that affect
that and we have no resolution.
-- Simon Riggs http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training & Services