Re: [HACKERS] GSoC 2017 : Patch for predicate locking in Gist index

From: Teodor Sigaev <teodor(at)sigaev(dot)ru>
To: Andrey Borodin <x4mmm(at)yandex-team(dot)ru>, Alexander Korotkov <a(dot)korotkov(at)postgrespro(dot)ru>
Cc: Shubham Barai <shubhambaraiss(at)gmail(dot)com>, Andrew Borodin <amborodin86(at)gmail(dot)com>, Heikki Linnakangas <hlinnaka(at)iki(dot)fi>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>, Kevin Grittner <kgrittn(at)gmail(dot)com>
Subject: Re: [HACKERS] GSoC 2017 : Patch for predicate locking in Gist index
Date: 2018-03-27 12:44:48
Message-ID: 1df13a56-8b13-d869-832e-0f52a2e1052e@sigaev.ru
Views: Raw Message | Whole Thread | Download mbox
Thread:
Lists: pgsql-hackers

Thanks to everyone, pushed

Andrey Borodin wrote:
>
>
>> 27 марта 2018 г., в 13:45, Alexander Korotkov <a(dot)korotkov(at)postgrespro(dot)ru
>> <mailto:a(dot)korotkov(at)postgrespro(dot)ru>> написал(а):
>>
>> On Tue, Mar 27, 2018 at 11:16 AM, Andrey Borodin <x4mmm(at)yandex-team(dot)ru
>> <mailto:x4mmm(at)yandex-team(dot)ru>> wrote:
>>
>> > 27 марта 2018 г., в 12:53, Teodor Sigaev <teodor(at)sigaev(dot)ru
>> <mailto:teodor(at)sigaev(dot)ru>> написал(а):
>> >
>> > I have a question: why do not CheckForSerializableConflictIn() move  into begining of gistplacetopage()? Seems, it is the single
>> function which actually changes page and all predicate locking stuff will
>> be placed in single function...
>>
>> gistplacetopage() is called from
>> 1. Buffered build - probably harmless
>>
>>
>> Yes, harmless, but useless.
>>
>> 2. Finish split - i'm not sure about this. It seems to me that it is
>> necessary... then your version is correct.
>>
>>
>> Yes, it's necessary, because GiST scan can end up on non-leaf page.  So, scan
>> and modify of same non-leaf page should conflict.
>>
>> Checking for serializable conflicts from buffering build seems useless
>> overhead.  gistplacetopage()
>> is called from only two places: gistinserttuples()
>> and gistbufferinginserttuples().  In order to evade
>> useless overhead for buffering build, I've moved
>> CheckForSerializableConflictIn() into gistinserttuples().
> +1
> Also, both gistdoinsert() and gistplacetopage() are mind blowing in complexity.
> gistinserttuples() looks like a cosy place if we need to add some more.
>
> Best regards, Andrey Borodin.

--
Teodor Sigaev E-mail: teodor(at)sigaev(dot)ru
WWW: http://www.sigaev.ru/

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Tomas Vondra 2018-03-27 12:54:44 Re: pg_class.reltuples of brin indexes
Previous Message Tomas Vondra 2018-03-27 12:23:55 Re: Online enabling of checksums