Horton

Re: GiST VACUUM

Thomas Munro <thomas.munro@enterprisedb.com>

View thread View on postgr.es
From: Thomas Munro <thomas.munro@enterprisedb.com>
To: Andrey Borodin <x4mmm@yandex-team.ru>
Cc: Heikki Linnakangas <hlinnaka@iki.fi>, Robert Haas <robertmhaas@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>, Костя Кузнецов <chapaev28@ya.ru>
Date: 2018-07-29T09:04:35Z
Lists: pgsql-hackers 
On Tue, Jul 24, 2018 at 6:04 AM, Andrey Borodin <x4mmm@yandex-team.ru> wrote:
>> 21 июля 2018 г., в 17:11, Andrey Borodin <x4mmm@yandex-team.ru> написал(а):
>> <0001-Physical-GiST-scan-in-VACUUM-v13.patch>
>
> Just in case, here's second part of patch series with actual page deletion.

Hi Andrey,

Cfbot says:

https://ci.appveyor.com/project/postgresql-cfbot/postgresql/build/1.0.7146

That's because you declared a new variable after some other
statements.  You can't do that in old school C89.

https://travis-ci.org/postgresql-cfbot/postgresql/builds/409401951

That segfaulted here:

#0 0x00000000004ab620 in setinternal (state=<optimized out>,
state=<optimized out>, blkno=368) at gistvacuum.c:44
44 state->internalPagesMap[blkno / 8] |= 1 << (blkno % 8);

internalPagesMap was NULL, or pointed to memory that was too small and
happened to be near an unmapped region (unlikely?), or was some other
corrupted address?

-- 
Thomas Munro
http://www.enterprisedb.com

Commits

  1. Use full 64-bit XID for checking if a deleted GiST page is old enough.

  2. Refactor checks for deleted GiST pages.

  3. Delete empty pages during GiST VACUUM.

  4. Scan GiST indexes in physical order during VACUUM.

  5. Prevent GIN deleted pages from being reclaimed too early