Mark Wong escribió:
> Hrm, tracking just the launcher process certainly doesn't help. Are
> the spawned processed short lived? I take a snapshot of
> /proc/<pid>/io data every 60 seconds.
The worker processes can be short-lived, but if they are, obviously they
are not vacuuming the large tables. If you want to track all autovacuum
actions, change autovacuum_log_min_messages to 0.
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
In response to
pgsql-performance by date
|Next:||From: Kynn Jones||Date: 2008-12-23 17:25:48|
|Subject: How to "unique-ify" HUGE table?|
|Previous:||From: Greg Smith||Date: 2008-12-23 03:51:08|
|Subject: Re: dbt-2 tuning results with postgresql-8.3.5|