Re: checkpoint_timeout irrelevant in 8.1.9

From: Alvaro Herrera <alvherre(at)commandprompt(dot)com>
To: George Wilk <gwilk(at)ellacoya(dot)com>
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: checkpoint_timeout irrelevant in 8.1.9
Date: 2007-06-20 17:56:29
Message-ID: 20070620175629.GO30369@alvh.no-ip.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

George Wilk wrote:
> According to the WAL config documentation correctly, checkpoints are created
> every
> <mk:@MSITStore:C:\Program%20Files\PostgreSQL\8.2\pgAdmin%20III\docs\en_US\pg
> admin3.chm::/pg/runtime-config-wal.html#guc-checkpoint-segments>
> checkpoint_segments log segments, or every
> <mk:@MSITStore:C:\Program%20Files\PostgreSQL\8.2\pgAdmin%20III\docs\en_US\pg
> admin3.chm::/pg/runtime-config-wal.html#guc-checkpoint-timeout>
> checkpoint_timeout seconds, whichever comes first. However, it seems as
> though setting my checkpoint_timeout has no effect on checkpoint creation (
> I have it set to 30 seconds). Forcing checkpoint by running psql -c
> CHECKPOINT mydb doesn't do the trick either. Checkpoints are created solely
> based on the checkpoint_segments logs reaching their configured size.

You are misunderstand what a checkpoint is. Also, why would you need a
checkpoint every 30 seconds escapes me. What you probably want is to
archive the WAL segment currently in use. There's a function you can
use for that, pg_switch_xlog().

--
Alvaro Herrera Valdivia, Chile ICBM: S 39º 49' 18.1", W 73º 13' 56.4"
"In Europe they call me Niklaus Wirth; in the US they call me Nickel's worth.
That's because in Europe they call me by name, and in the US by value!"

In response to

Browse pgsql-admin by date

  From Date Subject
Next Message Anastasios Hatzis 2007-06-20 17:57:58 Re: On managerial choosing (was: Postgres VS Oracle)
Previous Message Joshua D. Drake 2007-06-20 17:45:55 Re: checkpoint_timeout irrelevant in 8.1.9