Re: 7.3.4 on Linux: UPDATE .. foo=foo+1 degrades massivly
От | Guy Fraser |
---|---|
Тема | Re: 7.3.4 on Linux: UPDATE .. foo=foo+1 degrades massivly |
Дата | |
Msg-id | 408D6A01.3050908@incentre.net обсуждение исходный текст |
Ответ на | Re: 7.3.4 on Linux: UPDATE .. foo=foo+1 degrades massivly over time ("Dann Corbit" <DCorbit@connx.com>) |
Ответы |
Re: 7.3.4 on Linux: UPDATE .. foo=foo+1 degrades massivly
|
Список | pgsql-general |
Alvaro Herrera wrote: >On Mon, Apr 26, 2004 at 12:53:09PM -0600, Guy Fraser wrote: > > > >>Eventually the automatic purging of 'stale' data will be supported, >>but hopefully it will be configurable to allow 'time travel' when >>required, and allow for a reasonable time to be able to roll back >>transactions. >> >> > >Well, you are saying two different things here: to garbage-collect >automatically the database (rather than by manual VACUUMs), and to be >able to UNDO transactions, effectively going back in time. > >The former is likely to be supported in some not-too-distant future, >maybe hopefully the next release; the latter is not even planned, and in >the past it has been disregarded as too costly. Not implementation >time cost, mind you, but runtime cost. > > I realize that one you vacuum you can no longer time travel to before the vacuum. Although I never tried to use it, I thoughttime travel was a feature in PostGreSQL. My understanding of the time travel feature was to allow a query to be processedwith the data set as it was at a previous time. Since I did not have need for that feature fro any of the projects I have been involved in, I did not check to see how it worked, or followed it's development or demise as it maybe. Thank you for the update, I will not use time travel in further explanations of transactional integrity.:-)
В списке pgsql-general по дате отправления: