Re: Defaulting wal_sync_method to fdatasync on Linux for 9.1? - Mailing list pgsql-performance

From Andres Freund
Subject Re: Defaulting wal_sync_method to fdatasync on Linux for 9.1?
Date
Msg-id 201011010010.28724.andres@anarazel.de
Whole thread Raw
In response to Re: Defaulting wal_sync_method to fdatasync on Linux for 9.1?  (Greg Smith <greg@2ndquadrant.com>)
Responses Re: Defaulting wal_sync_method to fdatasync on Linux for 9.1?  (Greg Smith <greg@2ndquadrant.com>)
List pgsql-performance
On Sunday 31 October 2010 20:59:31 Greg Smith wrote:
> Writes only are sync'd out when you do a commit, or the database does a
> checkpoint.
Hm?  WAL is written out to disk after an the space provided by wal_buffers(def
8) * XLOG_BLCKSZ (def 8192) is used. The default is 64kb which you reach
pretty quickly - especially after a checkpoint. With O_D?SYNC that will
synchronously get written out during a normal XLogInsert if hits a page
boundary.
*Additionally* its gets written out at a commit if sync commit is not on.

Not having a real O_DSYNC on linux until recently makes it even more dubious
to have it as a default...


Andres

pgsql-performance by date:

Previous
From: Greg Smith
Date:
Subject: Re: Defaulting wal_sync_method to fdatasync on Linux for 9.1?
Next
From: Marti Raudsepp
Date:
Subject: Re: Defaulting wal_sync_method to fdatasync on Linux for 9.1?