Re: TOAST performance (was Re: [GENERAL] Delete Perform - Mailing list pgsql-hackers

From Mikheev, Vadim
Subject Re: TOAST performance (was Re: [GENERAL] Delete Perform
Date
Msg-id 3705826352029646A3E91C53F7189E32518459@sectorbase2.sectorbase.com
Whole thread Raw
Responses Re: TOAST performance (was Re: [GENERAL] Delete Perform  (Bruce Momjian <pgman@candle.pha.pa.us>)
List pgsql-hackers
> If "pre-page WAL write" means the value of the page before the current
> changes, then there is generally another reason for writing it out.

Bruce, "*pre*-page" is confusing - we write "after-change" page
image to WAL.

> When the system comes back up, we need to do a rollback on
> transaction B since it did not commit and we need the "pre-page"
> to know how to undo the change for B that got saved in step 6 above.

Brian, PGSQL still uses non-overwriting storage manager -
removing rows inserted by aborted transactions is not required.

Vadim


pgsql-hackers by date:

Previous
From: Brian Beuning
Date:
Subject: Re: TOAST performance (was Re: [GENERAL] Delete Performance)
Next
From: mlw
Date:
Subject: Integer aggregate and enum