Re: Cost Model

From: neto brpr <netobrpr(at)gmail(dot)com>
To: "David G(dot) Johnston" <david(dot)g(dot)johnston(at)gmail(dot)com>
Cc: pgsql-hackers(at)lists(dot)postgresql(dot)org
Subject: Re: Cost Model
Date: 2017-12-20 19:26:53
Message-ID: CALhhvc1fWnq+2acUC4MoAXwiY3Wzj3+c9OpDvXgwzaihPO2C8w@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

2017-12-20 16:37 GMT-02:00 David G. Johnston <david(dot)g(dot)johnston(at)gmail(dot)com>:

> On Wed, Dec 20, 2017 at 11:26 AM, neto brpr <netobrpr(at)gmail(dot)com> wrote:
>
>> Dear David
>> I have read documentation that you send, but it has only sequential page
>> cost and random page cost parameters. What I need, would be a model of
>> custo for Differentiate Read/Write (sequential and random), because in SSDs
>> the reads and writes have different costs. If you or someone knows a patch
>> or other solution, that allows you to configure individual parameters to:
>>
>> - Sequential reading page cost
>> - cost of the random reading page
>> - sequential recording page cost
>> - Random recording page cost
>>
>>
> ​Please don't top-post.
>

OK, sorry!

>
> OK...reading more closely I don't see how "recording/writing" costs are
> important in decision making. Either you have to write something, or you
> don't.​ If you do you updates pages in a buffer and generate WAL and then
> the system puts the data onto disk where they belong - the database itself
> doesn't care about that part and knowing how fast or slow it might happen
> would impact it behavior. So PostgreSQL provides read settings to be tuned
> so it can decide between index and table scans on the table in question.
>
> The system doesn't understand SSD or HDD but does understand tablespaces
> and I believe that many of these settings are able to be configured on a
> per-tablespace (or table?) basis.
>
>
About what you said, that some settings can be configured by Tablespace?
I have already seen this in IBM DB2, but in Postgresql as far as I know,
for example the Random_page_cost and Seq_page_cost parameters are
configured for the Integer Database and not for Tablespace, ok?.
You or someone can tell me if cost parameters can be configured by
Tablespace, this would be useful for me, thinking of a server that has
hybrid storage environment (HDD x SSDs), I could leave some in a tablespace
with HDD adapted settings and the same way for when I have an SSD disk.

Regards
Neto

David J.
>
>

<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
Livre
de vírus. www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>.
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message David G. Johnston 2017-12-20 19:32:03 Re: Cost Model
Previous Message Tom Lane 2017-12-20 19:26:25 Missed parallelism option in plpgsql?