On 2013-12-01 18:02:27 -0500, Tom Lane wrote:
> Andres Freund <andres@2ndquadrant.com> writes:
> > On 2013-12-01 17:15:31 -0500, Tom Lane wrote:
> >> Is there really a significant risk of clog access errors due to this bug?
> >> IIUC, the risk is that tuples in pages that vacuum skips due to being
> >> all-visible might not be frozen when intended.
>
> > Unfortunately it's not actually too hard to hit due to following part of the
> > code in vacuumlazy.c:
>
> > /*
> > * If we're not scanning the whole relation to guard against XID
> > * wraparound, it's OK to skip vacuuming a page. The next vacuum
> > * will clean it up.
> > */
>
> Ah. So it's only been *seriously* broken since commit bbb6e559c, ie 9.2.
Well, even before that crash recovery/replication didn't necessarily
preserve the hint bits. Even more so if somebody dared to set
full_page_writes=off.
I personally think full_page_writes=off should conflict with wal_level
!= minimal, btw, but I don't see much chance of gaining acceptance for
that.
Greetings,
Andres Freund
-- Andres Freund http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training &
Services