From: | Dushyanth <dushyanth(at)gmail(dot)com> |
---|---|
To: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: [Postgresql 8.2.3] autovacuum starting up even after disabling ? |
Date: | 2008-08-12 06:27:48 |
Message-ID: | loom.20080812T061408-93@post.gmane.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Hi Tom & Alvaro,
> Hey,
>
> > > They are all under 200 million
> >
> > Weird
> >
> > Could you fetch from pg_stat_activity the table it's processing, and its
> > pg_class row and that of its toast table (if any)?
>
> Sorry for the delay. Required details are at
> http://pastebin.com/pastebin.php?dl=fd699fbb
Did you guys have a chance to look at this ? I scheduled a VACUUM ANALYZE
VERBOSE on all databases to run daily and this has been running since few weeks
now.
Also note that the vacuum on my primary database is run like below to avoid
statement timeouts (set to 120000 in postgresql.conf).
> vacuum.sql
SET STATEMENT_TIMEOUT TO 0;
VACUUM ANALYSE verbose;
psql -U postgres -d dbname -f /path/to/vacuum.sql
For other databases, i do 'vacuumdb -zv dbname' .
The vacuum logs and the pgfouine vacuum reports dont show anything funny.
Autovacuum still starts up though. This process started on 09-08-2008 11.40 GMT
and ran till 10-08-2008 08:00 GMT. It also seems to be touching few other tables
apart from the tables i posted last about.
postgres 30430 0.0 0.1 2270284 60500 ? S Jul24 2:15
/usr/local/postgres/pgsql-8.2.3/bin/postgres -D
/usr/local/postgres/current/foundationdata -i
postgres 30437 5.0 0.0 57380 760 ? Ds Jul24 1196:04 \_ postgres:
logger process
postgres 31907 0.0 6.4 2271528 2129736 ? Ss Jul24 8:48 \_ postgres:
writer process
postgres 31908 0.0 0.0 58448 844 ? Ss Jul24 0:06 \_ postgres:
archiver process
postgres 31909 0.0 0.0 58448 812 ? Ss Jul24 0:00 \_ postgres:
stats collector process
postgres 7112 1.8 1.7 2291200 570796 ? Ss 11:40 0:05 \_ postgres:
autovacuum process foundation
I don't know what autovacuum is panicking about to warrant a force run.
Any pointers ?
TIA
Dushyanth
From | Date | Subject | |
---|---|---|---|
Next Message | Willy-Bas Loos | 2008-08-12 08:29:57 | differnt behaviour of NULL in an aggregate and with an operator |
Previous Message | Matthew Dennis | 2008-08-12 03:35:26 | different results based solely on existence of index (no, seriously) |