Re: Proposed patch: remove hard-coded limit MAX_ALLOCATED_DESCS

Поиск
Список
Период
Сортировка
От Cédric Villemain
Тема Re: Proposed patch: remove hard-coded limit MAX_ALLOCATED_DESCS
Дата
Msg-id 201306091021.13437.cedric@2ndquadrant.com
обсуждение исходный текст
Ответ на Re: Proposed patch: remove hard-coded limit MAX_ALLOCATED_DESCS  (Tom Lane <tgl@sss.pgh.pa.us>)
Список pgsql-hackers
Le samedi 8 juin 2013 23:27:16, Tom Lane a écrit :
> Cédric Villemain <cedric@2ndquadrant.com> writes:
> > I'm not sure of expected value of "max_safe_fds". Your patch now
> > initialize with 5 slots instead of 10, if max_safe_fds is large maybe it
> > is better to double the size each time we need instead of jumping
> > directly to the largest size ?
>
> I don't see the point particularly.  At the default value of
> max_files_per_process (1000), the patch can allocate at most 500 array
> elements, which on a 64-bit machine would probably be 16 bytes each
> or 8KB total.

The point was only if the original comment was still relevant. It seems it is
just not accurate anymore.
With patch I can union 492 csv tables instead of 32 with beta1.
--
Cédric Villemain +33 (0)6 20 30 22 52
http://2ndQuadrant.fr/
PostgreSQL: Support 24x7 - Développement, Expertise et Formation

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Craig Ringer
Дата:
Сообщение: Re: Bad error message on valuntil
Следующее
От: Heikki Linnakangas
Дата:
Сообщение: Re: pgbench: introduce a new automatic variable 'client_number'