Re: Rename max_parallel_degree?

Поиск
Список
Период
Сортировка
От Julien Rouhaud
Тема Re: Rename max_parallel_degree?
Дата
Msg-id 79f47dd2-819a-940a-5912-5ff643e87b54@dalibo.com
обсуждение исходный текст
Ответ на Re: Rename max_parallel_degree?  (Julien Rouhaud <julien.rouhaud@dalibo.com>)
Ответы Re: Rename max_parallel_degree?  (Amit Kapila <amit.kapila16@gmail.com>)
Список pgsql-hackers
On 15/06/2016 18:14, Julien Rouhaud wrote:
> On 15/06/2016 17:49, Robert Haas wrote:
>> On Tue, Jun 14, 2016 at 7:10 AM, Julien Rouhaud
>> <julien.rouhaud@dalibo.com> wrote:
>>>> I don't entirely like the new logic in
>>>> RegisterDynamicBackgroundWorker.
>>>
>>>> I wonder if we can't drive this off
>>>> of a couple of counters, instead of having the process registering the
>>>> background worker iterate over every slot. [...]
>>>>
>>
>> I think we should go that way.  Some day we might try to make the
>> process of finding a free slot more efficient than it is today; I'd
>> rather not double down on linear search.
>>
>
> Ok.
>
>> Are you going to update this patch?
>>
>
> Sure, I'll post a new patch ASAP.
>

Attached v4 implements the design you suggested, I hope everything's ok.

I didn't do anything for the statistics part, because I'm not sure
having the counters separately is useful (instead of just having the
current number of parallel workers), and if it's useful I'd find strange
to have these counters get reset at restart instead of being stored and
accumulated as other stats, and that's look too much for 9.6 material.
I'd be happy to work on this though.

--
Julien Rouhaud
http://dalibo.com - http://dalibo.org

Вложения

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Kevin Grittner
Дата:
Сообщение: Re: Re: [COMMITTERS] pgsql: Avoid extra locks in GetSnapshotData if old_snapshot_threshold <
Следующее
От: Robert Haas
Дата:
Сообщение: Re: Re: [COMMITTERS] pgsql: Avoid extra locks in GetSnapshotData if old_snapshot_threshold <