Hi Kevin
Well, you're right :-) But my use cases are un-specific "by design"
since I'm using FTS as a general purpose function.
So I still propose to enhance the planner too as Tom Lane and your
colleague suggest based on repeated similar complaints [1].
Yours, Stefan
[1] http://www.postgresql.org/message-id/CA+TgmoZgQBeu2KN305hwDS+aXW7YP0YN9vZwBsbWA8Unst+cew@mail.gmail.com
2013/7/29 Kevin Grittner <kgrittn@ymail.com>:
> Stefan Keller <sfkeller@gmail.com> wrote:
>
>> Finally, setting random_page_cost to 1 helps also - but I don't
>> like this setting neither.
>
> Well, you should learn to like whichever settings best model your
> actual costs given your level of caching and your workload. ;-)
> FWIW, I have found page costs less volatile and easier to tune
> with cpu_tuple_cost increased. I just always start by bumping
> that to 0.03.
>
> --
> Kevin Grittner
> EDB: http://www.enterprisedb.com
> The Enterprise PostgreSQL Company