Re: question regarding full_page_write - Mailing list pgsql-general

From Martín Marqués
Subject Re: question regarding full_page_write
Date
Msg-id CABeG9Lu5zJoNaCcR0M=jgP8A7b5Dbc70VOW_genTGnMiAr8EYw@mail.gmail.com
Whole thread Raw
In response to Re: question regarding full_page_write  (Greg Smith <greg@2ndquadrant.com>)
Responses Re: question regarding full_page_write
List pgsql-general
2011/2/17 Greg Smith <greg@2ndquadrant.com>:
> AI Rumman wrote:
>>
>> I can't clearly understand what FULL_PAGE_WRITE parameter is stand for.
>> Documentation suggest that If I make it OFF, then I have the chance for DB
>> crash.
>> Can anyone please tell me how it could be happened?
>
> The database writes to disk in 8K blocks.  If you can be sure that your disk
> drives and operating system will always write in 8K blocks, you can get a
> performance improvement from turning full_page_writes off.  But if you do
> that, and it turns out that when the power is interrupted your disk setup
> will actually do partial writes of less than 8K, your database can get
> corrupted.  Your system needs to ensure that when a write happens, either
> the whole thing goes to disk, or none of it does.

Sorry for the late reply, but I was investigating this option in
postgresql.conf and saw this mail.

My question regarding your answer is, why is it important for the
first page after a checkpoint and not on other page writes?


--
Martín Marqués
select 'martin.marques' || '@' || 'gmail.com'
DBA, Programador, Administrador

pgsql-general by date:

Previous
From: Gary Chambers
Date:
Subject: Re: securing the sql server ?
Next
From: Greg Smith
Date:
Subject: Re: question regarding full_page_write