"Jim C. Nasby" <[EMAIL PROTECTED]> writes:
> I think it'd be better to attack this problem from the "other side";
> namely looking at what's actually cached.

You can kiss goodbye to plan stability if you go that route... and
in any case I doubt the assumption that what's in shared buffers is
representative of what's in kernel cache.

                        regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 5: don't forget to increase your free space map settings

Reply via email to