From: | Julian Scarfe <julian(dot)scarfe(at)ntlworld(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | <pgsql-general(at)postgresql(dot)org> |
Subject: | Re: Vacuum verbose output? |
Date: | 2003-01-14 16:04:08 |
Message-ID: | BA49E578.1FEF5%julian.scarfe@ntlworld.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general pgsql-hackers |
On 14/1/03 15:42, "Tom Lane" <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> "Julian Scarfe" <julian(dot)scarfe(at)ntlworld(dot)com> writes:
>> The application is a database where the data in the tables are essentially
>> write-once-read-many, and a tuple stays active for about 3 hours before
>> being deleted/archived. With a regular, simple VACUUM, the number of Pages
>> increases steadily until a VACUUM FULL is performed.
>
> What do you have the FSM parameters in postgresql.conf set to? My guess
> is you need to increase 'em.
Currently set at the defaults, max_fsm_relations is 100, max_fsm_pages is
10000.
What are the implications/effects of increasing them?
BTW, one of my tables is much bigger, turning over at a similar rate (though
I'm much less concerned about performance for that one). Typical VACUUM
FULL looks like:
NOTICE: Pages 94279: Changed 315, reaped 53522, Empty 0, New 0; Tup
2753324: Vac 7907, Keep/VTL 0/0, UnUsed 277009, MinLen 135, MaxLen 1248;
Re-using: Free/Avail. Space 8569160/3380588; EndEmpty/Avail. Pages 0/5802.
Thanks
Julian
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2003-01-14 16:10:30 | Re: Vacuum verbose output? |
Previous Message | Robert Treat | 2003-01-14 15:54:29 | Re: Vacuum verbose output? |
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2003-01-14 16:10:30 | Re: Vacuum verbose output? |
Previous Message | Robert Treat | 2003-01-14 15:54:29 | Re: Vacuum verbose output? |