On 01-02-2016 21:13, Andres Freund wrote:
> is there any reason for the rather arbitrary and low checkpoint_timeout
> limit?
>
AFAICS the only reason is to run recover quickly. This setting is the
same value since day 1.
> A high timeout has the advantage that the total amount of full page
> writes reduces and, especially if the whole system fits into s_b, that
> the total amount of writes to disk is drastically reduced.
>
This statement could be added to documentation. Using this use case, I
want to propose raising the c_t upper limit to one day or even a week.
> I'm not sure what'd actually be a good upper limit. I'd be inclined to
> even go to as high as a week or so. A lot of our settings have
> upper/lower limits that aren't a good idea in general.
>
A week is an insane default value. However, I'm fine with 10 until 20
minutes (those are the most common values I use for c_t).
-- Euler Taveira Timbira - http://www.timbira.com.br/ PostgreSQL: Consultoria, Desenvolvimento,
Suporte24x7 e Treinamento