Re: Vacuum rate limit in KBps

From: Alvaro Herrera <alvherre(at)commandprompt(dot)com>
To: Simon Riggs <simon(at)2ndquadrant(dot)com>
Cc: Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>, Greg Smith <greg(at)2ndquadrant(dot)com>, Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Vacuum rate limit in KBps
Date: 2012-01-19 19:15:12
Message-ID: 1327000472-sup-5934@alvh.no-ip.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers


Excerpts from Simon Riggs's message of jue ene 19 16:05:36 -0300 2012:
> On Sun, Jan 15, 2012 at 9:17 PM, Heikki Linnakangas
> <heikki(dot)linnakangas(at)enterprisedb(dot)com> wrote:
>
> > I think it makes more sense to use the max read rate as the main knob,
> > rather than write rate. That's because the max read rate is higher than the
> > write rate, when you don't need to dirty pages. Or do you think saturating
> > the I/O system with writes is so much bigger a problem than read I/O that it
> > makes more sense to emphasize the writes?
>
> Yes, the writes are more important of the two.
>
> Too many writes at one time can overflow hardware caches, so things
> tend to get much worse beyond a certain point.
>
> Also, rate limiting writes means we rate limit WAL rate also which is
> very important.
>
> I'd like this to apply to large DDL, not just VACUUMs.

More generally, this can sometimes be useful in general queries as well.

--
Álvaro Herrera <alvherre(at)commandprompt(dot)com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Simon Riggs 2012-01-19 19:17:13 Re: WAL Restore process during recovery
Previous Message Andrew Dunstan 2012-01-19 19:10:00 Re: automating CF submissions (was xlog location arithmetic)