Re: vacuum and autovacuum - is it good to configure the threshold at TABLE LEVEL?

From: amul sul <sulamul(at)gmail(dot)com>
To: Muthiah Rajan <vgmonnet(at)gmail(dot)com>
Cc: PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: vacuum and autovacuum - is it good to configure the threshold at TABLE LEVEL?
Date: 2018-12-10 09:06:23
Message-ID: CAAJ_b97vYNoKdDzaV47a8HcY6z76=HU_svx4McVWMg_AbgcLtQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Apologize for delayed response.

On Mon, Dec 3, 2018 at 12:30 PM rajan <vgmonnet(at)gmail(dot)com> wrote:
>
> Hello Amul,
>
> I have a doubt. Please find below the details,
>
> 1. autovacuum_vacuum_threshold and autovacuum_analyse_threshold are set at
> 300
> 2. autovacuum_vacuum_scale_factor and autovacuum_analyse_scale_face are set
> at 300
> 3. there is table which has only 6 records
> 4. and the number of updates happened in the table is 944
> 5. now when I fetch data from pg_stat_user_table, i find that only
> last_autoanalyse column is updated with datetime and last_autovacuum column
> is not having a value
>
> does this mean only autoanalyse was executed for this table? But when I have
> both vacuum_threshold and analyse_threshold set at 300, how can only analyse
> run and vacuum does not?
>
That's right auto-vacuumed might not have triggered -- I think, that because of
autovacuum_vacuum_scale_factor setting and most of the time update does not
increase the table size.

Following question might not directly be related to your query but
just curious to know:
Does your table have indexes and by any chance your experiment
updating non-index column?

Regards,
Amul

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message rajan 2018-12-10 09:21:46 Re: vacuum and autovacuum - is it good to configure the threshold at TABLE LEVEL?
Previous Message Takahashi, Ryohei 2018-12-10 08:13:21 RE: Too many logs are written on Windows (LOG: could not reserve shared memory region (addr=%p) for child %p:)