Re: [HACKERS] HOT WIP Patch - version 2

On 2/20/07, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>
> "Pavan Deolasee" <pavan(dot)deolasee(at)gmail(dot)com> writes:
> > ... Yes. The HOT-updated status of the root and all intermediate
> > tuples is cleared and their respective ctid pointers are made
> > point to themselves.
>
> Doesn't that destroy the knowledge that they form a tuple chain?
> While it might be that no one cares any longer, it would seem more
> reasonable to leave 'em chained together.
>
>
I see your point, but as you mentioned do we really care ? The chain
needs to be broken so that the intermediate DEAD tuples can be
vacuumed. We can't vacuum them normally because they could
be a part of live HOT-update chain. Resetting the HOT-updated
status of the root tuple helps to mark the index entry LP_DELETE
once the entire HOT-update chain is dead.

Also, if we decide to reuse the heap-only tuples without even
vacuuming, breaking the chain is a better option since we then
guarantee no references to the heap-only DEAD tuples.

Thanks,
Pavan

--

EnterpriseDB http://www.enterprisedb.com