On Thu, Jun 29, 2006 at 14:27:30 +0200,
Martijn van Oosterhout <kleptog@svana.org> wrote:
> On Thu, Jun 29, 2006 at 01:21:19PM +0100, Simon Riggs wrote:
> > The issue is the difference between start of transaction and time when
> > the serializable snapshot is taken. Since BEGIN and other commands may
> > be issued as separate network requests it makes sense to defer taking
> > the snapshot until the first time it is needed. The transaction is still
> > serializable, just that the manual is worded slightly incorrectly with
> > regards the exact timing.
>
> I've always interpreted it as "there exists a serialised order for the
> transactions" but the database makes no guarentees about what it might
> be. I can't think of any real world case where you actually care about
> the order, just as long as one exists.
Postgres' serializable mode doesn't guaranty that. To get that effect you
may need to do some extra locking.