Re: bt_index_parent_check and concurrently build indexes

Поиск
Список
Период
Сортировка
От Donghang Lin
Тема Re: bt_index_parent_check and concurrently build indexes
Дата
Msg-id CAA=D8a2Q23miHJtHRDk_TSQKvd6oHk8wGpkQt99B=9yd-oqnfg@mail.gmail.com
обсуждение исходный текст
Ответ на Re: bt_index_parent_check and concurrently build indexes  ("Andrey M. Borodin" <x4mmm@yandex-team.ru>)
Список pgsql-hackers
Hi Michail 
 
> It turns out that bt_index_parent_check is not suitable for validating indexes built concurrently. 
Your finding is right on point! We recently used bt_index_parent_check to verify concurrently built indexes in a concurrent workload, bt_index_parent_check often gave such false positive error. 

- indexinfo->ii_Concurrent = !state->readonly;
+ indexinfo->ii_Concurrent = true;

One suggestion to this change is that we might need to update the amcheck doc to reflect that
"This consists of a “dummy” CREATE INDEX CONCURRENTLY operation" rather than "CREATE INDEX" operation.

Regards,
Donghang

В списке pgsql-hackers по дате отправления: