Re: Update on true serializable techniques in MVCC
От | Albe Laurenz |
---|---|
Тема | Re: Update on true serializable techniques in MVCC |
Дата | |
Msg-id | D960CB61B694CF459DCFB4B0128514C2039380D7@exadv11.host.magwien.gv.at обсуждение исходный текст |
Ответ на | Re: Update on true serializable techniques in MVCC (Robert Haas <robertmhaas@gmail.com>) |
Список | pgsql-hackers |
Robert Haas wrote: > > A predicate can include columns from an index plus others. > > Am I missing something? > > Hmm, interesting point. In that case you couldn't use the index to > enforce predicate locking under MVCC without disabling HOT. But there > will be other cases where that wouldn't help anyway - a predicate > could also include unindexed columns exclusively. For those, the > traditional approach (not the one discussed in this paper) probably > requires locking against any heap insert, or checking each new heap > insert against the constraint, or... something. If I understand that correctly > [...] by acquiring a shared lock on the next > row in order, as a scan is made to check whether rows match a predicate. > The scan might be through the data records or through an index I would say that in the case of a table scan, the whole table will be SILOCKed. I guess that's pretty much unavoidable if you want serializability. Yours, Laurenz Albe
В списке pgsql-hackers по дате отправления: