Re: [HACKERS] proposals for LLL, part 2 (locking)

Поиск
Список
Период
Сортировка
От Vadim Mikheev
Тема Re: [HACKERS] proposals for LLL, part 2 (locking)
Дата
Msg-id 35C13006.35F3B822@krs.ru
обсуждение исходный текст
Ответ на proposals for LLL, part 2 (locking)  (Vadim Mikheev <vadim@krs.ru>)
Список pgsql-hackers
> But I would like to add two internal lock modes due to VACUUM.
>
> 6. AccessShareLock     - acquired by each DML statement
>                          (INSERT, UPDATE, DELETE, SELECT)
>                          for the duration of statement.
>
> 7. AccessExclusiveLock - acquired by VACUUM
>
> : we can't vacuum a relation scanned by some other backend...
> (BTW, having these ones we get rid of pg_vlock file...)

On the other hand, we could use ObjExclusiveLock for vacuum -
vacuuming relations opened by running transaction is not
much usefull thing for now...

>
> And now yet two another locks for DDL statements
> (DROP/ALTER) for tables and indices only.
>
> 8. ObjShareLock     - acquired by heap/index open for the duration
>                       of transaction.
>
> 9. ObjExclusiveLock - acquired by DROP/ALTER.

Vadim

В списке pgsql-hackers по дате отправления:

Предыдущее
От: "Thomas G. Lockhart"
Дата:
Сообщение: Re: [INTERFACES] psqlodbc
Следующее
От: Bruce Momjian
Дата:
Сообщение: OR clause status report