Re: Autovacuum integration

"Matthew T. O'Connor" <matthew(at)zeut(dot)net> writes:
> Tom Lane wrote:
>> Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> writes:
>>> Ok. Do you think it's worth the trouble to have ALTER TABLE commands to
>>> change autovac parameters?
>>
>> I think we'll want them eventually, but I don't mind if 8.1 ships
>> without 'em.
>>
> It might be good to ship 8.1 without them since autovacuum might change
> a fair amount between 8.1 and 8.2 (FSM integration, better formula for
> vacuum decisions etc...)

Yeah, I was thinking the same thing: the more support there is, the
harder it will be to change the parameter set. Leaving it as a catalog
only will send a suitable message that this is still a prototype ...

> If memory serves from development discussion of the contrib version,
> isn't there an issue that while a shared relation need only be vacuumed
> in one database in order to reclaim space etc, the analyze statistics
> need to be updated in each database. Does that ring a bell or am I way
> off base here?

That's a fair point, but I'm not sure how much we care about maintaining
stats for the shared relations. These days the planner should get it
reasonably right about the relation size anyway, and the availability of
pg_statistic data is only a secondary concern; especially since we have
unique indexes on most of the columns people might search on, and the
presence of such an index also clues the planner.

regards, tom lane