Re: DSM segment handle generation in background workers

Поиск
Список
Период
Сортировка
От Tom Lane
Тема Re: DSM segment handle generation in background workers
Дата
Msg-id 17043.1539046409@sss.pgh.pa.us
обсуждение исходный текст
Ответ на Re: DSM segment handle generation in background workers  (Thomas Munro <thomas.munro@enterprisedb.com>)
Ответы Re: DSM segment handle generation in background workers
Re: DSM segment handle generation in background workers
Список pgsql-hackers
Thomas Munro <thomas.munro@enterprisedb.com> writes:
> On Mon, Oct 8, 2018 at 1:17 AM Thomas Munro
> <thomas.munro@enterprisedb.com> wrote:
>> That's because the bgworker startup path doesn't contain a call to
>> srandom(...distinguishing stuff...), unlike BackendRun().  I suppose
>> do_start_bgworker() could gain a similar call... or perhaps that call
>> should be moved into InitPostmasterChild().  If we put it in there
>> right after MyStartTime is assigned a new value, we could use the same
>> incantation that PostmasterMain() uses.

> Maybe something like this?

I think the bit with

#ifndef HAVE_STRONG_RANDOM
     random_seed = 0;
     random_start_time.tv_usec = 0;
#endif

should also be done in every postmaster child, no?  If we want to hide the
postmaster's state from child processes, we ought to hide it from all.

            regards, tom lane



В списке pgsql-hackers по дате отправления:

Предыдущее
От: "Iwata, Aya"
Дата:
Сообщение: RE: Function for listing archive_status directory
Следующее
От: Thomas Munro
Дата:
Сообщение: Re: pread() and pwrite()