Re: We need to log aborted autovacuums
Tom Lane <tgl@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Itagaki Takahiro <itagaki.takahiro@gmail.com>
Cc: Josh Berkus <josh@agliodbs.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>
Date: 2010-11-17T23:52:47Z
Lists: pgsql-hackers
Itagaki Takahiro <itagaki.takahiro@gmail.com> writes: > On Thu, Nov 18, 2010 at 08:35, Tom Lane <tgl@sss.pgh.pa.us> wrote: >> Well, the way to deal with that would be to add a GUC that enables >> reporting of those messages at LOG level. But it's a bit hard to argue >> that we need such a thing without more evidence. Maybe you could just >> locally modify the DEBUG3 to LOG and see whether it teaches you >> anything? > How about adding a special role for autovacuum, and running autovacuum > by the role instead of the database owner? If we have "autovacuum" role > for autovacuum processes, we could set log_min_messages to DEBUG3 for > only the role with per-user configuration. That seems like a major kluge ... and anyway it doesn't fix the problem, because DEBUG3 is still going to result in a lot of unwanted log output, even if it's confined to autovacuum. regards, tom lane