> Thomas,
>
> Thank you for your very detailed and well written description. In
> conclusion, I should keep my random_page_cost (3.0) to a value more than
> seq_page_cost (1.0)? Is this bad practice or will this suffice for my
> setup (where the database is much bigger than the RAM in the system)? Or
> is this not what you are suggesting at all?
Yes, keep it that way. The fact that 'random_page_cost >= seq_page_cost'
generally means that random reads are more expensive than sequential
reads. The actual values are dependent but 4:1 is usually OK, unless your
db fits into memory etc.
The decrease of performance after descreasing random_page_cost to 3 due to
changes of some execution plans (the index scan becomes slightly less
expensive than seq scan), but in your case it's a false assumption. So
keep it at 4 (you may even try to increase it, just to see if that
improves the performance).
regards
Tomas