From: | Simon Riggs <simon(at)2ndQuadrant(dot)com> |
---|---|
To: | Peter Geoghegan <pg(at)heroku(dot)com> |
Cc: | Pg Hackers <pgsql-hackers(at)postgresql(dot)org>, Magnus Hagander <magnus(at)hagander(dot)net> |
Subject: | Re: autovacuum_work_mem |
Date: | 2013-11-24 17:06:55 |
Message-ID: | CA+U5nMLyJx-ent1hm198QtVwsfDySZOqdK+8vAGZjh1LGj3dZA@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On 19 October 2013 19:22, Peter Geoghegan <pg(at)heroku(dot)com> wrote:
> I won't repeat the rationale for the patch here.
I can't see the problem that this patch is trying to solve. I'm having
trouble understanding when I would use this.
VACUUM uses 6 bytes per dead tuple. And autovacuum regularly removes
dead tuples, limiting their numbers.
In what circumstances will the memory usage from multiple concurrent
VACUUMs become a problem? In those circumstances, reducing
autovacuum_work_mem will cause more passes through indexes, dirtying
more pages and elongating the problem workload.
I agree that multiple concurrent VACUUMs could be a problem but this
doesn't solve that, it just makes things worse.
Freezing doesn't require any memory at all, so wraparound vacuums
won't be controlled by this parameter.
Can we re-state what problem actually is here and discuss how to solve
it. (The reference [2] didn't provide a detailed explanation of the
problem, only the reason why we want a separate parameter).
--
Simon Riggs http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
From | Date | Subject | |
---|---|---|---|
Next Message | Alexey Vasiliev | 2013-11-24 17:10:16 | Re[2]: [HACKERS] Connect from background worker thread to database |
Previous Message | Tom Lane | 2013-11-24 16:52:51 | Re: Re: Server is not getting started with log level as debug5 on master after commit 3147ac |