Re: Vacuums on large busy databases

From: Francisco Reyes <lists(at)stringsutils(dot)com>
To: Dave Cramer <pg(at)fastcrypt(dot)com>
Cc: Pgsql performance <pgsql-performance(at)postgresql(dot)org>
Subject: Re: Vacuums on large busy databases
Date: 2006-09-14 17:17:40
Message-ID: cone.1158254260.480661.31944.1000@zoraida.natserv.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

Dave Cramer writes:

> What is effective_cache set to ?

Default of 1000. Was just reading about this parameter.
Will try increasing it to 8192 (8192 * 8K = 64MB)

> why not just let autovac do it's thing ?

Have been playing with decresing the autovac values. With 100GB+ tables even
1% in autovacuum_vacuum_scale_factor is going to be 1GB.

Right now trying:
autovacuum_vacuum_threshold = 50000
autovacuum_analyze_threshold = 100000
autovacuum_vacuum_scale_factor = 0.05
autovacuum_analyze_scale_factor = 0.1

Initially I had tried autovacuum_vacuum_scale_factor = 0.2 and that was not
enough. Had to end up bumping fsm_pages several times. After I started to do
the 3 daily vacuums they are holding steady.. perhaps I will try only 1
vacuum now that I decreased the threshold to 0.05

I would be curious what others have in their autovacuum parameters for
100GB+ databases

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Francisco Reyes 2006-09-14 17:36:55 Re: Vacuums on large busy databases
Previous Message Dave Cramer 2006-09-14 16:57:46 Re: Vacuums on large busy databases