Re: PITR Phase 2 - Design Planning
Peter Galbavy <peter.galbavy@knowtion.net>
From: Peter Galbavy <peter.galbavy@knowtion.net>
To: Bruno Wolff III <bruno@wolff.to>
Cc: pgsql-hackers@postgresql.org
Date: 2004-04-28T08:28:02Z
Lists: pgsql-hackers
Bruno Wolff III wrote: > The context of my suggestion was for recovering up until a transaction which > messed things up was committed. I did not want the problem transaction to > be committed. If the problem transaction ran for a long time, there might > be other transactions that I want to keep, if possible, that committed > after the problem transaction started and before it ended. Ah! followed by Eek! Now I see the light. It's very bright and painful. What I can see is that expressing this accurately and unambiguously is going to be _difficult_. How do you know accurately the point just before a transaction was completed. There must be a good subset of candidates that can be labelled. Is there anyway to label/name a transaction that can be kept somewhere ? Like "begin transaction 'bigtrasacation26';" - is there any allowance in the SQL standards for naming trasactions ? PS I have fixed my system clock - apologies to my earlier reply being a month ahead. rgds, -- Peter