Re: Autovacuum Improvements

On 19/12/06, Chris Browne <cbbrowne(at)acm(dot)org> wrote:
> matthew(at)zeut(dot)net ("Matthew O'Connor") writes:
> > 2) Once we can have multiple autovacuum workers: Create the concept of
> > hot tables that require more attention and should never be ignored for
> > more that X minutes, perhaps have one "autovacuum worker" per hot
> > table? (What do people think of this?)
>
> One worker per "hot table" seems like overkill to me; you could chew
> up a lot of connections that way, which could be a DOS.

Sounds like a max workers config varible would work quite well here.
Bit like the max connections varible. If we run out of workers we just
have to wait for one to finish. I think we need one daemon to analyse
what needs vacuuming and then lauch workers to do the actual work..

Peter Childs

>
> That you have a "foot gun" is guaranteed; I think I'd rather that it
> come in the form that choosing the "hot list" badly hurts the rate of
> vacuuming than that we have a potential to chew up numbers of
> connections (which is a relatively non-renewable resource).
> --
> (format nil "~S(at)~S" "cbbrowne" "cbbrowne.com")
> http://linuxdatabases.info/info/
> There are no "civil aviation for dummies" books out there and most of
> you would probably be scared and spend a lot of your time looking up
> if there was one. :-) -- Jordan Hubbard in c.u.b.f.m
>
> ---------------------------(end of broadcast)---------------------------
> TIP 5: don't forget to increase your free space map settings
>