HEAD seems to generate larger WAL regarding GIN index

From: Fujii Masao <masao(dot)fujii(at)gmail(dot)com>
To: PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: HEAD seems to generate larger WAL regarding GIN index
Date: 2014-03-15 18:40:03
Message-ID: CAHGQGwE3xnytNBY0mnLbQ9+QvgPvL44iEE6bcnsm86eanPWEjg@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

I executed the following statements in HEAD and 9.3, and compared
the size of WAL which were generated by data insertion in GIN index.

---------------------
CREATE EXTENSION pg_trgm;
CREATE TABLE hoge (col1 text);
CREATE INDEX hogeidx ON hoge USING gin (col1 gin_trgm_ops) WITH
(FASTUPDATE = off);

CHECKPOINT;
SELECT pg_switch_xlog();
SELECT pg_switch_xlog();

SELECT pg_current_xlog_location();
INSERT INTO hoge SELECT 'POSTGRESQL' FROM generate_series(1, 1000000);
SELECT pg_current_xlog_location();
---------------------

The results of WAL size are

960 MB (9.3)
2113 MB (HEAD)

The WAL size in HEAD was more than two times bigger than that in 9.3.
Recently the source code of GIN index has been changed dramatically.
Is the increase in GIN-related WAL intentional or a bug?

Regards,

--
Fujii Masao

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Heikki Linnakangas 2014-03-15 19:27:22 Re: HEAD seems to generate larger WAL regarding GIN index
Previous Message jack 2014-03-15 16:50:11 BUG #9578: Undocumented behaviour for temp tables created inside query language (SQL) functions