Re: CLUSTER and MVCC

On 3/9/07, Florian G. Pflug <fgp(at)phlo(dot)org> wrote:
> Heikki Linnakangas wrote:
> > Csaba Nagy wrote:
> >> On Fri, 2007-03-09 at 14:00, Alvaro Herrera wrote:
> >>> But I'm not really seeing the problem here. Why isn't Csaba's problem
> >>> fixed by the fact that HOT reduces the number of dead tuples in the
> >>> first place? If it does, then he no longer needs the CLUSTER
> >>> workaround, or at least, he needs it to a much lesser extent.
> >>
> >> Is this actually true in the case of HOT + long running transactions ? I
> >> was supposing HOT has the same problems in the presence of long running
> >> transactions...
> >
> > It does, HOT won't help you here. A long-running transaction is just as
> > much of a problem with HOT as without. Besides, I don't recall that
> > you're doing updates in the first place.
>
> Couldn't HOT in principle deal with this? Let's say you have two long-running
> transactions, which see row versions A and D. While those transactions
> are running, the row is constantly updated, leading to row versions B, C (before
> the second long-running transaction started), D, E, F, ... Z.
> Now, the versions B,C,E,F,...Z could be removed by HOT or vacuum, because they
> are not currently visible, nor will they ever become visible because they are
> already deleted.

Couldn't they (or at least one of them) become visible due to
SAVEPOINT rollback?

>
> greetings, Florian Pflug
>
> ---------------------------(end of broadcast)---------------------------
> TIP 3: Have you checked our extensive FAQ?
>
> http://www.postgresql.org/docs/faq
>

--
Mike Rylander
mrylander(at)gmail(dot)com
GPLS -- PINES Development
Database Developer
http://open-ils.org