Re: GiST VACUUM

Heikki Linnakangas <hlinnaka@iki.fi>

From: Heikki Linnakangas <hlinnaka@iki.fi>

To: Andrey Borodin <x4mmm@yandex-team.ru>

Cc: Michael Paquier <michael@paquier.xyz>, Thomas Munro <thomas.munro@enterprisedb.com>, Robert Haas <robertmhaas@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>, Костя Кузнецов <chapaev28@ya.ru>

Date: 2019-03-22T12:03:59Z

Lists: pgsql-hackers

On 22/03/2019 13:37, Heikki Linnakangas wrote:
> On 21/03/2019 19:04, Heikki Linnakangas wrote:
>> Attached is the latest patch version, to be applied on top of the
>> IntegerSet patch.
> 
> And committed. Thanks Andrey!

This caused the buildfarm to go pink... I was able to reproduce it, by 
running the regression tests in one terminal, and repeatedly running 
"VACUUM;" in another terminal. Strange that it seems to happen all the 
time on the buildfarm, but never happened to me locally.

Anyway, I'm investigating.

- Heikki

Commits

Use full 64-bit XID for checking if a deleted GiST page is old enough.
- fb5344c969af 12.0 landed
- 6655a7299d83 13.0 landed
Refactor checks for deleted GiST pages.
- e2e992c93145 12.0 landed
- 9eb5607e6993 13.0 landed
Delete empty pages during GiST VACUUM.
- 7df159a620b7 12.0 landed
Scan GiST indexes in physical order during VACUUM.
- fe280694d0d4 12.0 landed
Prevent GIN deleted pages from being reclaimed too early
- 52ac6cd2d0cd 12.0 cited