Обсуждение: effective_io_concurrency in 9.6beta

Поиск
Список
Период
Сортировка

effective_io_concurrency in 9.6beta

От
Jeff Janes
Дата:
commit 1aba62ec made zero be an illegal value for effective_io_concurrency.

i think this was an accident.  If not, then the sample postgresql.conf
(at least) needs to be updated.

Cheers,

Jeff



Re: effective_io_concurrency in 9.6beta

От
Tom Lane
Дата:
Jeff Janes <jeff.janes@gmail.com> writes:
> commit 1aba62ec made zero be an illegal value for effective_io_concurrency.
> i think this was an accident.  If not, then the sample postgresql.conf
> (at least) needs to be updated.

It looks like the problem is that the new range check

+   /* This range check shouldn't fail, but let's be paranoid */
+   return (new_prefetch_pages > 0.0 && new_prefetch_pages < (double) INT_MAX);

should be testing for >= 0.0 not > 0.0.
        regards, tom lane



Re: effective_io_concurrency in 9.6beta

От
Alvaro Herrera
Дата:
Tom Lane wrote:
> Jeff Janes <jeff.janes@gmail.com> writes:
> > commit 1aba62ec made zero be an illegal value for effective_io_concurrency.
> > i think this was an accident.  If not, then the sample postgresql.conf
> > (at least) needs to be updated.
> 
> It looks like the problem is that the new range check
> 
> +   /* This range check shouldn't fail, but let's be paranoid */
> +   return (new_prefetch_pages > 0.0 && new_prefetch_pages < (double) INT_MAX);
> 
> should be testing for >= 0.0 not > 0.0.

Hmm, yeah, it looks like that's it.  Will fix.

-- 
Álvaro Herrera                http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services



Re: effective_io_concurrency in 9.6beta

От
Jeff Janes
Дата:
On Tue, May 24, 2016 at 11:34 AM, Alvaro Herrera
<alvherre@2ndquadrant.com> wrote:
> Tom Lane wrote:
>> Jeff Janes <jeff.janes@gmail.com> writes:
>> > commit 1aba62ec made zero be an illegal value for effective_io_concurrency.
>> > i think this was an accident.  If not, then the sample postgresql.conf
>> > (at least) needs to be updated.
>>
>> It looks like the problem is that the new range check
>>
>> +   /* This range check shouldn't fail, but let's be paranoid */
>> +   return (new_prefetch_pages > 0.0 && new_prefetch_pages < (double) INT_MAX);
>>
>> should be testing for >= 0.0 not > 0.0.
>
> Hmm, yeah, it looks like that's it.  Will fix.
>


Thanks, works as expected now.

Cheers,

Jeff