Re: speed up a logical replica setup

Поиск
Список
Период
Сортировка
От Euler Taveira
Тема Re: speed up a logical replica setup
Дата
Msg-id 6cc75dab-07bd-44bd-94a0-fe49ae550f34@app.fastmail.com
обсуждение исходный текст
Ответ на Re: speed up a logical replica setup  (Peter Eisentraut <peter@eisentraut.org>)
Список pgsql-hackers
On Mon, Jun 17, 2024, at 8:04 AM, Peter Eisentraut wrote:
On 07.06.24 05:49, Euler Taveira wrote:
> Here it is a patch series to fix the issues reported in recent 
> discussions. The
> patches 0001 and 0003 aim to fix the buildfarm issues. The patch 0002 
> removes
> synchronized failover slots on subscriber since it has no use. I also 
> included
> an optional patch 0004 that improves the usability by checking both 
> servers if
> it already failed in any subscriber check.

I have committed 0001, 0002, and 0003.  Let's keep an eye on the 
buildfarm to see if that stabilizes things.  So far it looks good.

Thanks.

For 0004, I suggest inverting the result values from check_publisher() 
and create_subscriber() so that it returns true if the check is ok.


Yeah, the order doesn't matter after the commit b963913826. I thought about
changing the order but I didn't; I provided a minimal patch. Since you think
it is an improvement, I'm attaching another patch with this change.


--
Euler Taveira

Вложения

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Peter Eisentraut
Дата:
Сообщение: replace strtok()
Следующее
От: Peter Eisentraut
Дата:
Сообщение: CompilerWarnings task does not catch C++ warnings