Re: the right time to vacuum database?

On Jun 8, 10:20 pm, a(dot)(dot)(dot)(at)crankycanuck(dot)ca (Andrew Sullivan) wrote:
> On Fri, Jun 08, 2007 at 06:29:55AM -0700, Charles.Hou wrote:
>
> > i have traced the size of the table. About 1000 new rows will be
> > inserted into the table in one day. each row has 300 bytes.
> > 1000*300/1024=293K, but the size of this table had been increased 3MB.
> > 3MB-293K=2.7MB...Why?where is the 2.7MB?
>
> How do you know what the size of the table is? You had at least two
> tables you were working on before. I think there must be something
> you're not communicating completely. (And why are you worried about
> less than 3 Meg anyway? Regular vacuum will leave some empty space
> around for new data, which means you don't have to go down to the
> filesystem to make the file bigger before you write it it. This is a
> Good Thing.)
>
> Please go back and run VACUUM VERBOSE on the table you killed the
> vacuum on before.
>
> A
>
> --
> Andrew Sullivan | a(dot)(dot)(dot)(at)crankycanuck(dot)ca
> However important originality may be in some fields, restraint and
> adherence to procedure emerge as the more significant virtues in a
> great many others. --Alain de Botton
>
> ---------------------------(end of broadcast)---------------------------
> TIP 2: Don't 'kill -9' the postmaster

because if the size of database have been increased 36MB in one day ,
then it will be 1G MB increased after 1 month. so i worry about the
size. other strange thing, if i block all postgresql client and run
vacuumdb, there will have about 100MB free space.

> How do you know what the size of the table is?
i got the relname from the pg_class and go to find the relname on the
disk.