Re: What is the relationship between checkpoint and wal

From: Kevin Grittner <kgrittn(at)ymail(dot)com>
To: 高健 <luckyjackgao(at)gmail(dot)com>, Jeff Janes <jeff(dot)janes(at)gmail(dot)com>
Cc: Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>, pgsql-general <pgsql-general(at)postgresql(dot)org>
Subject: Re: What is the relationship between checkpoint and wal
Date: 2013-08-30 20:25:21
Message-ID: 1377894321.70069.YahooMailNeo@web162906.mail.bf1.yahoo.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

高健 <luckyjackgao(at)gmail(dot)com> wrote:

> So I think that in a mission critical environment, it is not a
> good choice to turn full_page_writes on.

If full_page_writes is off, your database can be corrupted in the
event of a crash of the hardware, OS, or VM (for example a power
failure).  The only exception is if your environment somehow
guarantees that in such failures it is not possible to write part
of a 4KB write request without the entire 4KB being written.  Such
corruption may be hidden and result in inconsistent or incorrect
results, without generating an error; so you would be well-advised
to restore from backup if there is such a crash.

full_page_writes = on is required for protection of database
integrity in most environments.

In the mission critical environments I've worked with, it has
generally been desirable to preserve database integrity and to be
able to recover from an OS crash faster than can be done from
backup.

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Kevin Grittner 2013-08-30 20:33:09 Re: virtualxid,relation lock
Previous Message Tom Lane 2013-08-30 17:47:02 Re: Unable to CREATE SCHEMA and INSERT data in table in that schema in same EXECUTE