Thread: Effective Cache Size

Effective Cache Size

From
Darcy Buskermolen
Date:
I was thinking of changing the way EFC is handled, having it retrived via code 
(much the same way top does it) rather than relying on the user to specify 
the value.  This value would then be checked/updated every so often.

Does anybody have any thoughts on this.  Especially on if this would be a boon 
or a bust to the planner/optimizer.



-- 
Darcy Buskermolen
Wavefire Technologies Corp.
ph: 250.717.0200
fx:  250.763.1759
http://www.wavefire.com


Re: Effective Cache Size

From
Tom Lane
Date:
Darcy Buskermolen <darcy@wavefire.com> writes:
> I was thinking of changing the way EFC is handled, having it retrived via code 
> (much the same way top does it) rather than relying on the user to specify 
> the value.  This value would then be checked/updated every so often.

Even assuming that you can portably find out how much RAM is going to
kernel cache (which I do not believe), how do you know how much of it
is getting used for Postgres data as opposed to everything else?

At the moment I think that manual handling is sufficient anyway.  This
is a relatively low-impact parameter and it doesn't seem too important
to try to get it dead on.
        regards, tom lane