From: | Alvaro Herrera <alvherre(at)commandprompt(dot)com> |
---|---|
To: | Magnus Hagander <magnus(at)hagander(dot)net> |
Cc: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Greg Stark <greg(dot)stark(at)enterprisedb(dot)com>, PG Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: maintenance memory vs autovac |
Date: | 2008-12-03 12:58:28 |
Message-ID: | 20081203125828.GC3968@alvh.no-ip.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Magnus Hagander wrote:
> Tom Lane wrote:
> > It seems like mostly a confusion-generator to me. Is there any actual
> > evidence that autovac should use a different maintenance_work_mem than
> > other processes?
>
> The use-case that made me think of that is one with lots of autovac
> workers in a system with lots of small tables in different databases.
Another thing to consider here is that lazy vacuum will scale down its
memory usage depending on table size.
> Turns out I read the documentation for autovac wrong. I understood that
> if I wanted it to look at 1000 databases at once, I needed
> autovac_workers at 1000. Talked a bit offlist with Alvaro and realized
> that's not what it is, but that the documentation is a bit unclear on
> that - will work on fixing that.
Yeah, Rob Treat has also asked me twice about this, so it's probably
worth rewriting.
--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.
From | Date | Subject | |
---|---|---|---|
Next Message | ohp | 2008-12-03 13:13:01 | Re: cvs head initdb hangs on unixware |
Previous Message | Alvaro Herrera | 2008-12-03 12:52:32 | Re: snapshot leak and core dump with serializable transactions |