Re: Vacuum and FSM page size - Mailing list pgsql-performance

From Thomas Lozza
Subject Re: Vacuum and FSM page size
Date
Msg-id F6B04D920899DF47B1E1A4F40415167E3D4370@gaia.nexus-ag.com
Whole thread Raw
In response to Re: Vacuum and FSM page size  ("Scott Marlowe" <scott.marlowe@gmail.com>)
Responses Re: Vacuum and FSM page size
List pgsql-performance
Thanks for the advice.
I used the default settings before, thought though that vacuum was a bit
aggressive, ie, using too many resources. Now its taking very long. So
will have to find reasonable settings in between I guess.

On the other hand, if I keep the fsm_page number high enough, the system
should be fine with a low number of vacuum cycles, right. As memory is
not really scarce (16G, 32 bit PG though) an x million fsm_page entry
should be ok. Any thoughts on that?

cheers,
-- tom.



-----Original Message-----
From: Scott Marlowe [mailto:scott.marlowe@gmail.com]
Sent: Thursday, 24. January, 2008 10:48
To: Thomas Lozza
Cc: pgsql-performance@postgresql.org
Subject: Re: [PERFORM] Vacuum and FSM page size

On Jan 23, 2008 12:29 PM, Thomas Lozza <thomas.lozza@nexustelecom.com>
wrote:
> Auto vacuum is enabled. Here are the settings:
>
> autovacuum = true
> autovacuum_naptime = 900
> autovacuum_vacuum_threshold = 2000
> autovacuum_analyze_threshold = 1000
> autovacuum_vacuum_scale_factor = 0.25
> autovacuum_analyze_scale_factor = 0.18 autovacuum_vacuum_cost_delay =
> 150 autovacuum_vacuum_cost_limit = 120
>
> A manual vacuum takes very long (around 4 days), so maybe the cost
> delay and limit or too high.

Your autovacuum_vacuum_cost_delay is REALLY high.  Try setting it to 10
or 20 and see if that helps.

What is your plain old vacuum_cost_delay set to?

pgsql-performance by date:

Previous
From: "Merlin Moncure"
Date:
Subject: Re: Slow set-returning functions
Next
From: "Scott Marlowe"
Date:
Subject: Re: Vacuum and FSM page size