Re: [HACKERS] Parallel Index Scans
| От | Robert Haas |
|---|---|
| Тема | Re: [HACKERS] Parallel Index Scans |
| Дата | |
| Msg-id | CA+TgmoZxdAT72VAG90A-RKOFLLCOwL9xEbz-aOOhM4VhjmBxzw@mail.gmail.com обсуждение исходный текст |
| Ответ на | Re: [HACKERS] Parallel Index Scans (Amit Kapila <amit.kapila16@gmail.com>) |
| Ответы |
Re: [HACKERS] Parallel Index Scans
|
| Список | pgsql-hackers |
On Mon, Feb 13, 2017 at 9:04 PM, Amit Kapila <amit.kapila16@gmail.com> wrote: > I think the comment at that place is not as clear as it should be. So > how about changing it as below: > > Existing comment: > -------------------------- > /* > * For parallel scans, get the last page scanned as it is quite > * possible that by the time we try to fetch previous page, other > * worker has also decided to scan that previous page. So we > * can't rely on _bt_walk_left call. > */ > > Modified comment: > -------------------------- > /* > * For parallel scans, it is quite possible that by the time we try to fetch > * the previous page, another worker has also decided to scan that > * previous page. So to avoid that we need to get the last page scanned > * from shared scan descriptor before calling _bt_walk_left. > */ That sounds way better. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company
В списке pgsql-hackers по дате отправления: