Re: Snapshot backups

From: Bruce Momjian <bruce(at)momjian(dot)us>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: Alban Hertroys <haramrae(at)gmail(dot)com>, Jeff Janes <jeff(dot)janes(at)gmail(dot)com>, James Sewell <james(dot)sewell(at)lisasoft(dot)com>, "pgsql-general(at)postgresql(dot)org" <pgsql-general(at)postgresql(dot)org>
Subject: Re: Snapshot backups
Date: 2013-08-09 22:39:07
Message-ID: 20130809223907.GA30558@momjian.us
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Wed, Jul 31, 2013 at 08:24:46AM -0400, Tom Lane wrote:
> Alban Hertroys <haramrae(at)gmail(dot)com> writes:
> > That begs the question what happens in case of a crash or (worse) a partial crash when multiple file systems are involved.
>
> As long as the OS+hardware honors the contract of fsync(), everything's
> fine. If the storage system loses data that it claims to have fsync'd to
> stable storage, there's not much we can do about that, except recommend
> that you have a backup plan.
>
> In practice, the more complicated your storage infrastructure is, the more
> likely it is to have bugs ...

Just to give the 10k mile answer, the WAL contains all database changes
that _might_ be lost due to file system changes during a base backup.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Bruce Momjian 2013-08-09 23:26:42 Re: Type to to_char(d, 'J')?
Previous Message Tom Lane 2013-08-09 22:13:44 Re: plpgsql FOR LOOP CTE problem ?