Re: Talking about optimizer, my long dream

From: Cédric Villemain <cedric(dot)villemain(dot)debian(at)gmail(dot)com>
To: mladen(dot)gogala(at)vmsinfo(dot)com
Cc: Frank Heikens <frankheikens(at)mac(dot)com>, tivv00(at)gmail(dot)com, pgsql-performance <pgsql-performance(at)postgresql(dot)org>
Subject: Re: Talking about optimizer, my long dream
Date: 2011-02-04 15:45:39
Message-ID: AANLkTimCOWDC++T9e473S-VjxGd14HRSBPosnOrA8a55@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

2011/2/4 Frank Heikens <frankheikens(at)mac(dot)com>:
>
>
> On 04 Feb, 2011,at 02:56 PM, Mladen Gogala <mladen(dot)gogala(at)vmsinfo(dot)com>
> wrote:
>
> Віталій Тимчишин wrote:
>> Hi, all.
>>
>> All this optimizer vs hint thread
> There is no "optimizer vs. hint". Hints are a necessary part of the
> optimizer in all other databases.
>
>
> That has nothing to do with PostgreSQL: PostgreSQL = PostgreSQL. And it
> doesn't have hints and everybody knows it.
>
> Without hints Postgres will not get
> used in the company that I work for, period.
>
>
> That's up to you, that's fine. But why did you start with PostgreSQL in the
> first place? You knew PostgreSQL doesn't have hints and the wiki told you
> hints are not wanted as well. When hints are an essential requirement for
> your company, you should pick another product, EnterpriseDB Postgres Plus
> for example.
>
> I was willing to wait but
> the fatwa against hints seems unyielding,
>
>
> There is no fatwa. The PostgreSQL project prefers to spend resources on a
> better optimizer to solve the real problems, not on hints for working around
> the problems. That has nothing to do with any fatwa or religion.
>
> so that's it. I am even
> inclined to believe that deep down under the hood, this fatwa has an
> ulterior motive, which disgusts me deeply. With hints, there would be
> far fewer consulting gigs.
>
>
> The consulting guys are the ones who love hints: They know they have to come
> back the other month because the old hint does more harm than good when data
> changes. And data will change over time.
>
> You said it's so simple to implement hints in PostgreSQL, so please, show
> us. Or ask/pay somebody to write this simple code for you to support hints,
> nobody will ever stop you from doing that. When you have a use case that
> proves the usage of hints will improve the performance of PostgreSQL and you
> have some code that can be maintained by the PostgreSQL project, it might be
> implemented in the contrib or even core. It's up to you, not somebody else.

Just in case you miss it:
http://www.sai.msu.su/~megera/wiki/plantuner

Btw feel free to do how you want, it is open source, and BSD, you can
take PostgreSQL, add hints, go and sell that to your boss.

>
>
>
> Mladen Gogala
> Sr. Oracle DBA
> 1500 Broadway
> New York, NY 10036
> (212) 329-5251
> www.vmsinfo.com
>
>
> --
> Sent via pgsql-performance mailing list (pgsql-performance(at)postgresql(dot)org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-performance
>

--
Cédric Villemain               2ndQuadrant
http://2ndQuadrant.fr/     PostgreSQL : Expertise, Formation et Support

In response to

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Scott Marlowe 2011-02-04 16:05:32 Re: [HACKERS] Slow count(*) again...
Previous Message Tom Lane 2011-02-04 15:41:04 Re: Does auto-analyze work on dirty writes? (was: Re: [HACKERS] Slow count(*) again...)