Date: Fri, 8 Nov 2024 11:44:24 -0600
From: Nathan Bossart <nathandbossart@gmail.com>
To: wenhui qiu <qiuwenhuifx@gmail.com>
Cc: =?iso-8859-1?Q?Fr=E9d=E9ric?= Yhuel <frederic.yhuel@dalibo.com>,
	Robert Haas <robertmhaas@gmail.com>,
	"Imseih (AWS), Sami" <simseih@amazon.com>,
	David Rowley <dgrowleyml@gmail.com>,
	Joe Conway <mail@joeconway.com>, Michael Banck <mbanck@gmx.net>,
	Laurenz Albe <laurenz.albe@cybertec.at>,
	Melanie Plageman <melanieplageman@gmail.com>,
	PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: New GUC autovacuum_max_threshold ?
Message-ID: <Zy5N-M2jPIYzPdku@nathan>
References: <20240507211702.GA2720371@nathanxps13>
 <0A799687-DE0A-4BFC-8900-E1D57EB6C211@amazon.com>
 <CA+TgmoYF3HqEUArbsqtCtP7YMJ62G3U8W1mSmZFZoCK8uBSi9g@mail.gmail.com>
 <cc23c226-e0be-47a7-bf6f-bcedd097a239@dalibo.com>
 <CA+TgmoaER+O40PA_CB8eSKm7KT37wDT+-y_mnsWXb9-Keivu1Q@mail.gmail.com>
 <ZnD509Q1jROYKhz_@nathan>
 <97f04e09-a08a-417c-b65d-5d5cbdd0b845@dalibo.com>
 <ZrPprRLBzi1YUBLn@nathan>
 <6a2ac9b7-6535-4bb1-8274-0647f7c31c82@dalibo.com>
 <CAGjGUALGtQ8MoshY6tEUPiAOjWNefBzq61gZ_5Dx-Uai1cC39Q@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
In-Reply-To: <CAGjGUALGtQ8MoshY6tEUPiAOjWNefBzq61gZ_5Dx-Uai1cC39Q@mail.gmail.com>
Archived-At: <https://www.postgresql.org/message-id/Zy5N-M2jPIYzPdku%40nathan>
Precedence: bulk

On Wed, Nov 06, 2024 at 08:51:07PM +0800, wenhui qiu wrote:
>> Thank you. FWIW, I would prefer a sub-linear growth, so maybe something
>> like this
> 
>>   vacthresh = Min(vac_base_thresh + vac_scale_factor * reltuples,
>>   vac_base_thresh + vac_scale_factor * pow(reltuples, 0.7) * 100);
> 
>>   This would give :
> 
>>   * 386M (instead of 5.1 billion currently) for a 25.6 billion tuples
> table ;
>>   * 77M for a 2.56 billion tuples table (Robert's example) ;
>>   * 15M (instead of 51M currently) for a 256M tuples table ;
>>   * 3M (instead of 5M currently) for a 25.6M tuples table.
>> The other advantage is that you don't need another GUC.
> Argee ,We just need to change the calculation formula,But I prefer this
> formula because it calculates a smoother value.
> 
> vacthresh =  (float4) fmin(vac_base_thresh + vac_scale_factor *
> reltuples,vac_base_thresh
> + vac_scale_factor * log2(reltuples) * 10000);
> or
> vacthresh = (float4) fmin(vac_base_thresh + (vac_scale_factor * reltuples)
> , sqrt(1000.0 * reltuples));

I apologize for the curt response, but I don't understand how we could
decide which of these three complicated formulas to use, let alone how we
could expect users to reason about the behavior.

-- 
nathan