Re: crash-safe visibility map, take three

On Tue, Nov 30, 2010 at 11:55 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> writes:
>> Can we get away with not setting the LSN on the heap page, even though
>> we set the PD_ALL_VISIBLE flag? If we don't set the LSN, the heap page
>> can be flushed to disk before the WAL record, but I think that's fine
>> because it's OK to have the flag set in the heap page even if the VM bit
>> is not set.
>
> Why is that fine?  It's certainly not fine from the standpoint of
> someone wondering why his index-only scan performs so badly.
>
> I think all this hair-splitting about cases where it's okay to have one
> bit set and not the other is misguided.  To me, crash-safety of the VM
> means that its copy of the page-header bit is right.  Period.  Yes, it
> will cost something to ensure that; so what?  If we don't get more than
> enough compensating performance gain from index-only scans, the whole
> patch is going to end up reverted.

We're not going to double the cost of VACUUM to get index-only scans.
And that's exactly what will happen if you do full-page writes of
every heap page to set a single bit.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company