Re: tuning autovacuum

Em 08-06-2011 20:35, Robert Haas escreveu:
> Is the hint correct? I mean, what if there were 100 small tables that
> needed vacuuming all at the same time. We'd hit this limit no matter
> how high you set autovacuum_max_workers, but it wouldn't be right to
> set it to 101 just because every once in a blue moon you might trip
> over the limit.
>
I think so. You are picturing a scene with only one message. It is the same
case of the too-frequent-checkpoint messages; i.e., you should look if those
messages have some periodicity.

> I think it'd be really useful to expose some more data in this area
> though. One random idea is - remember the time at which a table was
> first observed to need vacuuming. Clear the timestamp when it gets
> vacuumed. Then you can do:
>
Hmmm. But this fine grained information alone doesn't help tuning the number
of autovacuum workers. I consider counters easier to implement and simpler to
analyze. But the timestamp idea has its merit because we already have a
similar statistic (last timestamp table was vacuumed or analyzed).

--
Euler Taveira de Oliveira - Timbira http://www.timbira.com.br/
PostgreSQL: Consultoria, Desenvolvimento, Suporte 24x7 e Treinamento